Skip to content

Hypergraph is more powerful. Transform unstructured text into structured knowledge with LLMs. Graphs, hypergraphs, and spatio-temporal extractions — with one command.

feat(template-optimizer): 新增模板优化技能及参考规则

yifanfeng97e2b8d48276cb
diff --git a/hyperextract-skills/README.md b/hyperextract-skills/README.md
index 97b6f23..938a414 100644
--- a/hyperextract-skills/README.md
+++ b/hyperextract-skills/README.md
@@ -12,6 +12,7 @@ Knowledge template design skills for Hyper-Extract.
| graph-designer | Design graph/hypergraph/etc structures |
| yaml-validator | Validate YAML configurations |
| multilingual | Convert to multi-language support |
+| template-optimizer | Optimize templates and fix common issues |
## Installation
@@ -43,8 +44,9 @@ cp -r hyperextract-skills ~/.trae/skills/
1. Start with the brainstorm skill to explore your requirements
2. Based on the recommended type, use either record-designer or graph-designer
-3. Optionally validate with yaml-validator
-4. Optionally convert to multilingual
+3. **Optimize with template-optimizer** (recommended)
+4. Optionally validate with yaml-validator
+5. Optionally convert to multilingual
### Example Workflow
@@ -111,8 +113,15 @@ hyperextract-skills/
│ ├── rules-types.md
│ ├── rules-identifiers.md
│ └── rules-errors.md
-└── multilingual/
- └── SKILL.md # Multi-language conversion
+├── multilingual/
+│ └── SKILL.md # Multi-language conversion
+└── template-optimizer/
+ ├── SKILL.md # Template optimization
+ └── references/ # Optimization rules
+ ├── rules-naming.md
+ ├── rules-multilingual.md
+ ├── rules-field-count.md
+ └── rules-consistency.md
```
## Design Principles
diff --git a/hyperextract-skills/graph-designer/SKILL.md b/hyperextract-skills/graph-designer/SKILL.md
index 22cd263..049e11b 100644
--- a/hyperextract-skills/graph-designer/SKILL.md
+++ b/hyperextract-skills/graph-designer/SKILL.md
@@ -204,3 +204,21 @@ identifiers:
- [ ] relation_id template correct?
- [ ] relation_members configured?
- [ ] time/location_field specified?
+
+### Multi-language
+- [ ] `zh` descriptions use pure Chinese (no English terms like `entity(...)`)
+- [ ] `en` descriptions use pure English (no Chinese characters)
+
+### Entity-Relation Consistency
+- [ ] source/target types match entities definition
+
+### Relation Type Design
+For **generic templates**: Use open-ended description
+```yaml
+en: 'Relation type, concisely describing the connection between entities'
+```
+
+For **domain-specific templates**: Can use predefined types
+```yaml
+en: 'Relation type: type_a/type_b/type_c'
+```
diff --git a/hyperextract-skills/multilingual/SKILL.md b/hyperextract-skills/multilingual/SKILL.md
index b3ff782..7813c25 100644
--- a/hyperextract-skills/multilingual/SKILL.md
+++ b/hyperextract-skills/multilingual/SKILL.md
@@ -94,6 +94,17 @@ rules:
- 'Each field corresponds to an independent information item'
```
+### Multi-language Consistency
+
+When translating field values (like enum options), maintain language purity:
+
+| Original | Chinese Translation | English Translation |
+|----------|-------------------|-------------------|
+| `类型:实体/抽象` | `类型:实体/抽象` | `Type: entity/abstract` |
+| `重要性:高/中/低` | `重要性:高/中/低` | `Significance: high/medium/low` |
+
+Do NOT translate inline: `类型:entity(实体)` → wrong!
+
## Translation Checklist
- [ ] Preserve technical terms consistently
diff --git a/hyperextract-skills/record-designer/SKILL.md b/hyperextract-skills/record-designer/SKILL.md
index f26614a..c6444d4 100644
--- a/hyperextract-skills/record-designer/SKILL.md
+++ b/hyperextract-skills/record-designer/SKILL.md
@@ -131,3 +131,7 @@ identifiers:
- [ ] Quality requirements specified?
- [ ] Common mistakes warned?
- [ ] For set: deduplication rules clear?
+
+### Multi-language
+- [ ] `zh` descriptions use pure Chinese
+- [ ] `en` descriptions use pure English
diff --git a/hyperextract-skills/template-optimizer/SKILL.md b/hyperextract-skills/template-optimizer/SKILL.md
new file mode 100644
index 0000000..8eed8c5
--- /dev/null
+++ b/hyperextract-skills/template-optimizer/SKILL.md
@@ -0,0 +1,174 @@
+---
+name: template-optimizer
+description: |
+ Optimize YAML templates for Hyper-Extract.
+ Use when: "optimize template", "fix YAML issues", "improve quality", "lint template"
+ Trigger: After creating templates or during review
+ Skip: Creating new templates (use brainstorm + designer instead)
+---
+
+# Template Optimizer
+
+Automatically analyze and optimize YAML templates by applying best practices and fixing common issues.
+
+## Workflow
+
+```
+1. Parse YAML → 2. Analyze Issues → 3. Apply Fixes → 4. Generate Report
+```
+
+### Step 1: Parse YAML
+
+Load the YAML template and validate basic structure.
+
+### Step 2: Analyze Issues
+
+Check against these rules:
+
+| Rule | What to Check |
+|------|--------------|
+| [rules-naming.md](references/rules-naming.md) | Field naming standardization |
+| [rules-multilingual.md](references/rules-multilingual.md) | Language consistency |
+| [rules-field-count.md](references/rules-field-count.md) | Information density |
+| [rules-consistency.md](references/rules-consistency.md) | Schema vs Guideline separation |
+
+### Step 3: Apply Fixes
+
+Apply fixes based on optimization level:
+
+| Level | Description | Example |
+|-------|-------------|---------|
+| **Auto-fix** | Safe changes that always improve | `relation_type` → `type` |
+| **Suggest** | Changes that may need review | Field count > 5 |
+| **Review** | Design decisions | Relation type openness |
+
+### Step 4: Generate Report
+
+Output changes with explanations for learning.
+
+---
+
+## Detection Rules
+
+### Rule 1: Multi-language Consistency
+
+```
+❌ Pattern: [a-zA-Z]+\([^)]+\) in zh fields
+❌ Pattern: Chinese chars in en fields
+```
+
+**Fix**: Separate language content, use pure Chinese in zh, pure English in en.
+
+### Rule 2: Field Naming
+
+```
+❌ relation_type → type
+❌ event_date → time
+❌ entity_type → type
+```
+
+**Fix**: Standardize to concise names.
+
+### Rule 3: Field Count
+
+```
+⚠️ entities.fields > 5 → Flag for review
+⚠️ relations.fields > 5 → Flag for review
+```
+
+**Fix**: Simplify to essential fields, use priority: Essential → Important → Optional.
+
+### Rule 4: Schema-Guideline Separation
+
+```
+❌ Repetition of field definitions in guideline
+❌ Schema descriptions in rules
+```
+
+**Fix**: Schema defines WHAT, Guideline defines HOW TO DO WELL.
+
+---
+
+## Design Principles
+
+### Schema vs Guideline
+
+| Schema Defines | Guideline Defines |
+|---------------|-----------------|
+| Field names | Extraction strategy |
+| Field types | Quality requirements |
+| Field descriptions | Creation conditions |
+| Required/optional | Common mistakes |
+
+**❌ Wrong**: Guideline repeats schema definitions
+**✅ Correct**: Guideline explains how to extract well
+
+### Information Density
+
+For **entities** and **relations**:
+
+| Field Priority | Examples |
+|---------------|---------|
+| Essential | source, target, participants |
+| Important | type, time, location |
+| Optional | description, metadata |
+
+**Max**: 5 fields per component
+
+### Naming Conventions
+
+| For | Use | Example |
+|-----|------|---------|
+| Template name | CamelCase | `EarningsSummary` |
+| Field names | snake_case | `company_name` |
+| Tags | lowercase | `finance, investor` |
+
+---
+
+## Reference Files
+
+| Topic | When to Check |
+|-------|--------------|
+| [rules-naming.md](references/rules-naming.md) | Field naming issues |
+| [rules-multilingual.md](references/rules-multilingual.md) | Language consistency |
+| [rules-field-count.md](references/rules-field-count.md) | Too many fields |
+| [rules-consistency.md](references/rules-consistency.md) | Schema-guideline overlap |
+
+---
+
+## Output Format
+
+```markdown
+# Optimization Report
+
+## Changes Made
+
+| File | Issue | Fix | Level |
+|------|-------|-----|-------|
+| template.yaml | `relation_type` | renamed to `type` | Auto-fix |
+| template.yaml | Mixed language | Fixed zh/en separation | Auto-fix |
+| template.yaml | 7 relation fields | Suggested simplification | Suggest |
+
+## Summary
+
+- Auto-fix: 2
+- Suggestions: 1
+- Manual review: 0
+```
+
+---
+
+## Integration
+
+```
+Full workflow: brainstorm → designer → optimizer → validator
+ ↓
+ Apply best practices
+ Auto-fix common issues
+```
+
+**When to use**:
+- After creating new templates
+- Before validator
+- During template review
+- Batch optimization of existing templates
diff --git a/hyperextract-skills/template-optimizer/references/rules-consistency.md b/hyperextract-skills/template-optimizer/references/rules-consistency.md
new file mode 100644
index 0000000..43f79b2
--- /dev/null
+++ b/hyperextract-skills/template-optimizer/references/rules-consistency.md
@@ -0,0 +1,122 @@
+# Schema vs Guideline Consistency
+
+Separation of concerns between schema definition and extraction guidelines.
+
+---
+
+## Core Principle
+
+| Schema (output) | Guideline (rules) |
+|-----------------|------------------|
+| Defines WHAT | Defines HOW TO DO WELL |
+| Field structure | Extraction strategy |
+| Type definitions | Quality requirements |
+| Descriptions | Common mistakes |
+
+---
+
+## What Belongs Where
+
+### Schema (output)
+
+```yaml
+output:
+ entities:
+ fields:
+ - name: name
+ type: str
+ description: 'Person name' # What it is
+ required: true
+```
+
+### Guideline (rules)
+
+```yaml
+guideline:
+ rules_for_entities:
+ - 'Extract full names as they appear in text'
+ - 'Maintain naming consistency throughout document'
+ - 'Only extract named persons, not pronouns'
+```
+
+---
+
+## Common Mistakes
+
+### ❌ Repeating Schema in Guideline
+
+```yaml
+# Wrong - repeats schema
+guideline:
+ rules_for_entities:
+ - 'Extract name field as entity name'
+ - 'Extract type field as entity category'
+```
+
+### ✅ Guideline Focuses on Extraction
+
+```yaml
+# Correct - explains how to extract
+guideline:
+ rules_for_entities:
+ - 'Extract full names as they appear'
+ - 'Categorize based on context clues'
+```
+
+---
+
+## Checklist
+
+| Check | Guideline Should | Guideline Should NOT |
+|-------|-----------------|---------------------|
+| Field names | Extraction strategy | "name should be..." |
+| Field types | Quality requirements | "type field is..." |
+| Required/optional | Creation conditions | "required: true" |
+| Descriptions | Common mistakes | Schema definitions |
+
+---
+
+## Examples
+
+### Model Template
+
+```yaml
+# Schema
+output:
+ fields:
+ - name: company_name
+ type: str
+ description: 'Company name'
+
+# Guideline
+guideline:
+ rules:
+ - 'Extract official company names'
+ - 'Use abbreviations if that's how they appear'
+ - 'Standardize formats like "Inc.", "Ltd."'
+```
+
+### Graph Template
+
+```yaml
+# Schema
+relations:
+ fields:
+ - name: type
+ description: 'Relationship type'
+
+# Guideline
+rules_for_relations:
+ - 'Use verbs or prepositions to describe relations'
+ - 'Match terminology from source text'
+ - 'Keep relation types consistent'
+```
+
+---
+
+## Detection
+
+Flag for review when guideline contains:
+- Field definitions from schema
+- "should be" or "must be" referencing schema fields
+- Repetition of `description` content
diff --git a/hyperextract-skills/template-optimizer/references/rules-field-count.md b/hyperextract-skills/template-optimizer/references/rules-field-count.md
new file mode 100644
index 0000000..0cc4600
--- /dev/null
+++ b/hyperextract-skills/template-optimizer/references/rules-field-count.md
@@ -0,0 +1,94 @@
+# Information Density Rules
+
+Guidelines for keeping templates focused and readable.
+
+---
+
+## Core Principle
+
+**Limit entity/relation fields to ≤5 for visual clarity and cognitive load.**
+
+| Component | Max Fields | Reason |
+|-----------|-----------|--------|
+| `entities.fields` | 5 | Visual clarity in graph nodes |
+| `relations.fields` | 5 | Readable edge labels |
+
+---
+
+## Field Priority
+
+### Essential (Always Include)
+
+| For | Fields |
+|-----|--------|
+| Graph | `source`, `target` |
+| Hypergraph | `participants` (or grouping fields) |
+| Temporal | `time` |
+| Spatial | `location` |
+
+### Important (Include If Needed)
+
+| For | Fields |
+|-----|--------|
+| Type identification | `type` |
+| Context | `description` |
+| Metadata | `category`, `role` |
+
+### Optional (Use Sparingly)
+
+| For | Fields |
+|-----|--------|
+| Additional context | `notes` |
+| Quality indicators | `confidence` |
+| External references | `source` |
+
+---
+
+## Simplification Strategies
+
+When fields exceed 5:
+
+1. **Merge related fields**: `first_name` + `last_name` → `name`
+2. **Move to description**: Move optional details to `description` field
+3. **Use grouping**: Extract sub-fields into nested structure
+4. **Prioritize**: Keep essential, mark others as optional
+
+---
+
+## Examples
+
+### ❌ Too Many Fields (7)
+
+```yaml
+relations:
+ fields:
+ - name: source
+ - name: target
+ - name: type
+ - name: strength
+ - name: description
+ - name: source_context
+ - name: target_context
+```
+
+### ✅ Simplified (5)
+
+```yaml
+relations:
+ fields:
+ - name: source
+ - name: target
+ - name: type
+ - name: strength
+ - name: description # Includes context info
+```
+
+---
+
+## Detection
+
+Flag for review when:
+- `entities.fields` > 5
+- `relations.fields` > 5
+
+**Note**: This is a suggestion, not an error. Some templates may legitimately need more fields.
diff --git a/hyperextract-skills/template-optimizer/references/rules-multilingual.md b/hyperextract-skills/template-optimizer/references/rules-multilingual.md
new file mode 100644
index 0000000..2c21a71
--- /dev/null
+++ b/hyperextract-skills/template-optimizer/references/rules-multilingual.md
@@ -0,0 +1,87 @@
+# Multi-language Consistency Rules
+
+Ensure language purity in bilingual templates.
+
+---
+
+## Core Principle
+
+**Each language field should use that language's own terminology.**
+
+| Field | Rule | Example |
+|-------|------|---------|
+| `zh` | Pure Chinese | `类型:实体` |
+| `en` | Pure English | `Type: entity` |
+
+---
+
+## Common Issues
+
+### Issue 1: Mixed Language in zh
+
+```
+❌ Wrong
+zh: '类型:entity(实体)/abstract(抽象概念)'
+en: 'Type: entity/abstract'
+
+✅ Correct
+zh: '类型:实体/抽象概念'
+en: 'Type: entity/abstract'
+```
+
+**Pattern to detect**: `[a-zA-Z]+\([^)]+\)` in zh fields
+
+### Issue 2: Mixed Language in en
+
+```
+❌ Wrong
+zh: '类型:实体/抽象'
+en: 'Type: 实体/抽象'
+
+✅ Correct
+zh: '类型:实体/抽象'
+en: 'Type: entity/abstract'
+```
+
+**Pattern to detect**: `[\u4e00-\u9fa5]+` in en fields
+
+---
+
+## Auto-fix Rules
+
+1. **Remove English terms from zh**: Extract `entity(...)` → `实体`
+2. **Remove Chinese chars from en**: Translate to English equivalents
+
+---
+
+## Translation Patterns
+
+When translating enum values:
+
+| Original (zh) | English (en) |
+|--------------|--------------|
+| 实体 | entity |
+| 抽象概念 | abstract |
+| 过程 | process |
+| 关系 | relation |
+| 高/中/低 | high/medium/low |
+| 人物/事件/地点 | person/event/location |
+
+---
+
+## Special Cases
+
+### Field Names vs Field Values
+
+```
+# Field names can be in English (they're identifiers)
+name: type
+description:
+ zh: '类型'
+ en: 'Type'
+
+# Field values should match language
+description:
+ zh: '类型:实体' # zh values in Chinese
+ en: 'Type: entity' # en values in English
+```
diff --git a/hyperextract-skills/template-optimizer/references/rules-naming.md b/hyperextract-skills/template-optimizer/references/rules-naming.md
new file mode 100644
index 0000000..968aec5
--- /dev/null
+++ b/hyperextract-skills/template-optimizer/references/rules-naming.md
@@ -0,0 +1,64 @@
+# Field Naming Standards
+
+Standardized field names for consistency across templates.
+
+---
+
+## Naming Conventions
+
+| Element | Convention | Example |
+|---------|-----------|---------|
+| Template name | CamelCase | `EarningsSummary` |
+| Field names | snake_case | `company_name` |
+| Tags | lowercase, comma-separated | `finance, investor` |
+
+---
+
+## Standard Field Names
+
+### Graph Types (relations)
+
+| Wrong | Correct | Reason |
+|-------|---------|--------|
+| `relation_type` | `type` | Redundant - already in relation context |
+| `event_date` | `time` | More generic for temporal graphs |
+| `location_info` | `location` | Shorter, clearer |
+| `participant_list` | `participants` | Consistent with hypergraph |
+
+### Entity Types
+
+| Wrong | Correct | Reason |
+|-------|---------|--------|
+| `entity_type` | `type` | Consistency with relations |
+| `object_type` | `type` | Shorter, same meaning |
+| `node_category` | `type` | Shorter |
+
+### Common Fields
+
+| Wrong | Correct | Reason |
+|-------|---------|--------|
+| `created_at` | `created` | Shorter |
+| `updated_at` | `updated` | Shorter |
+| `description_text` | `description` | Redundant suffix |
+
+---
+
+## Detection Patterns
+
+```
+relation_type → type
+entity_type → type
+event_date → time
+```
+
+---
+
+## Auto-fix Rules
+
+These are safe to auto-fix:
+
+1. `relation_type` → `type`
+2. `event_date` → `time`
+3. `entity_type` → `type`
+
+**Note**: Also update identifiers.relation_id template and display.relation_label references.
diff --git a/hyperextract-skills/yaml-validator/references/rules-syntax.md b/hyperextract-skills/yaml-validator/references/rules-syntax.md
index 7930f3d..8d0339d 100644
--- a/hyperextract-skills/yaml-validator/references/rules-syntax.md
+++ b/hyperextract-skills/yaml-validator/references/rules-syntax.md
@@ -82,3 +82,14 @@ options: {...}
2. Verify all required fields present
3. Check naming conventions
4. Validate field value types
+
+---
+
+## Multi-language Consistency
+
+| Field | Rule | ❌ Wrong | ✅ Correct |
+|-------|------|----------|------------|
+| `zh` | Pure Chinese | `类型:entity(实体)` | `类型:实体/抽象` |
+| `en` | Pure English | `Type: 实体` | `Type: entity/abstract` |
+
+Detection pattern: `[a-zA-Z]+\([^)]+\)` in `zh` fields.
diff --git a/hyperextract/cli/commands/list.py b/hyperextract/cli/commands/list.py
index e1a729c..96346b2 100644
--- a/hyperextract/cli/commands/list.py
+++ b/hyperextract/cli/commands/list.py
@@ -3,6 +3,7 @@
from typing import Optional
from rich.console import Console
from rich.table import Table
+from rich.text import Text
import typer
from hyperextract.utils.template_engine import Gallery
@@ -15,11 +16,11 @@ app = typer.Typer(
)
-def _get_description(cfg) -> str:
+def _get_description(cfg, lang: str = "en") -> str:
"""Get description from template config, handling both str and dict types."""
desc = cfg.description
if isinstance(desc, dict):
- return desc.get("zh", desc.get("en", ""))
+ return desc.get(lang, desc.get("en", ""))
return desc or ""
@@ -27,31 +28,36 @@ def _get_description(cfg) -> str:
def template(
query: Optional[str] = typer.Option(None, "--query", "-q", help="Query to search templates"),
autotype: Optional[str] = typer.Option(None, "--autotype", "-a", help="Filter by autotype"),
- language: Optional[str] = typer.Option(None, "--lang", "-l", help="Filter by language"),
+ language: Optional[str] = typer.Option(None, "--lang", "-l", help="Language filter (en/zh/all, default: en)"),
include_methods: bool = typer.Option(True, "--include-methods/--no-methods", help="Include method templates"),
):
"""List all available templates with detailed information."""
+ target_lang = language if language else "en"
+
results = Gallery.list(
filter_by_query=query,
filter_by_type=autotype,
- filter_by_language=language,
+ filter_by_language=None if target_lang == "all" else target_lang,
)
templates = []
for path, cfg in results.items():
- full_name = cfg.name
- languages = cfg.language if cfg.language else ["zh"]
+ languages = cfg.language if cfg.language else ["en"]
if isinstance(languages, str):
languages = [languages]
- for lang in languages:
- templates.append((path, full_name, lang, cfg.type, _get_description(cfg)))
+ if target_lang == "all":
+ for lang in languages:
+ templates.append((path, cfg.type, _get_description(cfg, lang)))
+ else:
+ lang = target_lang if target_lang in languages else (languages[0] if languages else "en")
+ templates.append((path, cfg.type, _get_description(cfg, lang)))
- if include_methods:
+ if include_methods and target_lang != "zh":
from hyperextract.methods import list_method_cfgs
method_templates = list_method_cfgs()
for name, cfg in method_templates.items():
- templates.append((name, cfg.name, "en", cfg.type, cfg.description))
+ templates.append((name, cfg.type, cfg.description))
if not templates:
console.print("[yellow]No templates found.[/yellow]")
@@ -62,16 +68,15 @@ def template(
show_header=True,
header_style="bold magenta",
expand=True,
+ show_lines=True,
)
- table.add_column("Template ID", style="cyan", no_wrap=True, width=30)
- table.add_column("Name", style="green", width=20)
- table.add_column("Type", style="yellow", no_wrap=True, width=12)
- table.add_column("Language", style="magenta", no_wrap=True, width=8)
+ table.add_column("Template ID", style="cyan", width=26)
+ table.add_column("Type", style="yellow", no_wrap=True, width=22)
table.add_column("Description", style="white")
- for tid, name, lang, template_type, desc in templates:
- desc_text = desc[:60] + "..." if len(desc) > 60 else desc
- table.add_row(tid, name, template_type, lang, desc_text)
+ for tid, template_type, desc in templates:
+ domain, name = tid.split("/", 1)
+ table.add_row(Text(f"{domain}/\n{name}", style="cyan"), template_type, desc)
console.print(table)
console.print(f"\n[dim]Total: {len(templates)} templates[/dim]")
diff --git a/hyperextract/templates/PLANNING_TEMPLATES_REDESIGN.md b/hyperextract/templates/PLANNING_TEMPLATES_REDESIGN.md
new file mode 100644
index 0000000..50942d6
--- /dev/null
+++ b/hyperextract/templates/PLANNING_TEMPLATES_REDESIGN.md
@@ -0,0 +1,879 @@
+# Hyper-Extract 知识模板重构计划
+
+## 📋 项目现状分析
+
+### 当前结构
+```
+templates/
+├── presets/ # 6个核心领域
+│ ├── finance/ # 金融
+│ ├── general/ # 通用(需保留base_*)
+│ ├── industry/ # 工业
+│ ├── legal/ # 法律
+│ ├── medicine/ # 医学
+│ └── tcm/ # 中医
+└── legacy/ # 旧模板(归档,待清理)
+ ├── presets/ # 6个归档领域
+ └── zh/ # Python模板
+```
+
+### 现有问题
+1. **过度设计**:创建了自我想象的需求
+2. **模板膨胀**:许多模板功能重叠,用户难以选择
+3. **脱离实际需求**:缺乏对技术趋势的响应
+
+---
+
+## 🔥 GitHub热点调研结论(2026年3月)
+
+### 关键技术趋势
+- **Agentic AI**:Sub-agents、Memory、Skills组合成为主流
+- **MCP (Model Context Protocol)**:正在成为Agent间通信标准
+- **Skills Marketplace**:3.7万+个可安装skill
+- **Workflow as Code**:YAML配置文件定义AI工作流
+
+---
+
+## 🎯 重新设计原则
+
+1. **用户真实需求驱动**:不是"可能有用",而是"真正需要"
+2. **控制模板数量**:每个领域5-8个精选模板
+3. **言简意赅命名**:模板名称简洁明了
+4. **每个图需描述**:明确用于什么文档、建模什么、可视化什么
+5. **保持base_*不变**:通用基础模板已完善
+
+---
+
+## 📚 1. **general** - 通用领域 ⭐
+
+### 设计原则
+通用领域服务于所有文档类型,重点关注AI时代新需求和通用文档分析。
+
+### 模板清单
+
+#### 保留项(9个)
+
+| 模板 | 类型 | 描述 |
+|------|------|------|
+| `base_graph` | graph | 通用图基础 |
+| `base_hypergraph` | hypergraph | 通用超图基础 |
+| `base_list` | list | 通用列表基础 |
+| `base_model` | model | 通用模型基础 |
+| `base_set` | set | 通用集合基础 |
+| `base_temporal_graph` | temporal_graph | 时序图基础 |
+| `base_spatial_graph` | spatial_graph | 空间图基础 |
+| `base_spatio_temporal_graph` | spatio_temporal_graph | 时空图基础 |
+
+#### 新增项(3个)
+
+##### 1.1 `workflow_graph` - 流程图
+
+**用于文档**:Skill定义文件(SKILL.md)、Agent工作流文档、SOP操作手册、业务流程说明
+
+**建模内容**:步骤名称、步骤类型(skill/step/condition/branch)、输入要求、输出结果、条件分支
+
+**可视化什么**:展示步骤之间的执行顺序、依赖关系、分支汇聚,形成流程可视化图
+
+```yaml
+type: temporal_graph
+核心要素:
+ - step_name: 步骤名称
+ - step_type: 类型(skill/step/condition/branch)
+ - description: 步骤描述
+ - input: 输入要求
+ - output: 输出结果
+关系:
+ - next: 顺序执行
+ - triggers: 触发条件
+ - branches_to: 条件分支
+ - converges: 汇聚点
+```
+
+---
+
+##### 1.2 `doc_structure` - 文档结构图
+
+**用于文档**:技术文档(README、API docs)、论文、报告、规范文档
+
+**建模内容**:章节标题、节点类型(chapter/section/paragraph/code/table)、层级深度、内容摘要、交叉引用
+
+**可视化什么**:展示文档的层级结构、章节关系、引用关系,便于理解长文档脉络
+
+```yaml
+type: graph
+核心要素:
+ - node_type: 节点类型(chapter/section/paragraph/code/table)
+ - title: 标题/名称
+ - level: 层级深度
+ - summary: 内容摘要
+关系:
+ - contains: 包含关系(父→子)
+ - references: 引用关系(交叉引用)
+```
+
+---
+
+##### 1.3 `biography_graph` - 传记图
+
+**用于文档**:人物传记、回忆录、年谱、人物介绍文章
+
+**建模内容**:事件名称、发生日期、发生地点、重要性、涉及人物、人物关系
+
+**可视化什么**:展示人物的关键事件人生轨迹、事件因果关系、人物互动网络
+
+```yaml
+type: temporal_graph
+核心要素:
+ - event: 事件名称
+ - event_date: 发生日期
+ - location: 发生地点
+ - significance: 重要性(高/中/低)
+ - participants[]: 涉及人物
+关系:
+ - precedes: 时间先后
+ - causes: 因果关系
+ - involves: 参与关系
+```
+
+---
+
+##### 1.4 `concept_graph` - 概念关系图
+
+**用于文档**:教科书、百科全书、学术论文、概念解释文档
+
+**建模内容**:概念名称、概念类型(实体/抽象/过程)、定义、层级关系
+
+**可视化什么**:展示概念之间的层级关系(is-a/part-of)、语义关联,形成知识地图
+
+```yaml
+type: graph
+核心要素:
+ - concept: 概念名称
+ - category: 概念类型(实体/抽象/过程/关系)
+ - definition: 简要定义
+关系:
+ - subclass_of: 层级关系(子类)
+ - part_of: 组成关系(部分)
+ - related_to: 相关关系
+```
+
+---
+
+### 删除项
+- ❌ `clause_list` → 合并到legal
+- ❌ `compliance_logic` → 合并到legal
+- ❌ `knowledge_graph` → 用base_graph替代
+- ❌ `social_network` → 过于泛化
+- ❌ `cross_reference_net` → 用doc_structure替代
+- ❌ `life_event_timeline` → 重命名为biography_graph
+
+**general预期数量:13个(原14个)**
+
+---
+
+## 💰 2. **finance** - 金融
+
+### 设计原则
+金融领域用户需求明确,保留高频使用模板。
+
+### 模板清单(5个)
+
+##### 2.1 `earnings_summary` - 财报会议摘要
+
+**用于文档**:财报电话会议记录(Earnings Call Transcript)
+
+**建模内容**:公司名、季度、收入/EPS实际值、与预期对比、前瞻指引、整体基调、亮点/关切
+
+**可视化什么**:快速理解会议核心内容,把握管理层信心和关键议题
+
+```yaml
+type: model
+核心字段:
+ - company_name, quarter
+ - reported_revenue, reported_eps
+ - revenue_vs_consensus, eps_vs_consensus
+ - guidance_revenue, guidance_eps
+ - overall_tone
+ - key_highlights[], key_concerns[]
+```
+
+---
+
+##### 2.2 `sentiment_model` - 市场情绪模型
+
+**用于文档**:财经新闻、市场评论、研究报告、社交媒体帖子
+
+**建模内容**:情绪分数(-1到1)、主导主题、关键实体、价格影响
+
+**可视化什么**:量化市场情绪变化,识别舆情与股价的关联
+
+```yaml
+type: model
+核心字段:
+ - sentiment_score: -1到1
+ - dominant_theme: 主导主题
+ - key_entities[]: 关键实体
+ - price_impact: 价格影响
+```
+
+---
+
+##### 2.3 `event_timeline` - 重大事件时间线
+
+**用于文档**:8-K报告、新闻稿、公司公告、投资评级调整
+
+**建模内容**:事件名称、事件日期、重大性(是否重大)、影响程度
+
+**可视化什么**:展示公司重要事件的时间顺序和因果关系,支持事件驱动分析
+
+```yaml
+type: temporal_graph
+核心要素:
+ - event: 事件名称
+ - event_date: 事件日期
+ - materiality: 重大性
+关系:
+ - causes: 因果关系
+ - precedes: 时间先后
+```
+
+---
+
+##### 2.4 `ownership_graph` - 股权结构图
+
+**用于文档**:招股说明书(IPO S-1)、年度报告(10-K)、股东名册
+
+**建模内容**:股东名称、持股比例、股东类型(机构/个人/关联方)、控制权关系
+
+**可视化什么**:展示股权层级结构、控制关系路径,辅助尽职调查和UBO识别
+
+```yaml
+type: graph
+核心要素:
+ - shareholder: 股东名称
+ - shareholder_type: 股东类型
+ - ownership_percentage: 持股比例
+关系:
+ - controls: 控制关系
+ - belongs_to: 归属集团
+```
+
+---
+
+##### 2.5 `risk_factor_set` - 风险因子集合
+
+**用于文档**:年报风险章节(10-K Item 1A)、季报风险披露、招股说明书风险章节
+
+**建模内容**:风险因子、风险类别(经营/财务/市场/法规)、严重程度
+
+**可视化什么**:提取和归类风险因子,便于风险监控和比较分析
+
+```yaml
+type: set
+核心要素:
+ - risk_factor: 风险因子
+ - severity: 严重程度
+ - category: 风险类别
+```
+
+---
+
+### 删除项
+- ❌ `SupplyChainGraph` → 合并到其他或删除
+- ❌ `MDANarrativeGraph` → 合并到event_timeline
+- ❌ `DiscussionGraph` → 使用率低
+- ❌ `FactorInfluenceHypergraph` → 过于复杂
+- ❌ `ValuationLogicMap` → 过于主观
+- ❌ 其他20+个模板 → 删除
+
+**finance预期数量:5个(原20+个)**
+
+---
+
+## 🏥 3. **medicine** - 医学
+
+### 设计原则
+医学文档标准化程度高,选择最通用的模板。
+
+### 模板清单(5个)
+
+##### 3.1 `treatment_map` - 治疗方案图
+
+**用于文档**:临床实践指南、治疗方案文档、临床路径文件
+
+**建模内容**:患者特征、诊断、治疗方案、适应症、禁忌组合、治疗结果
+
+**可视化什么**:展示诊断→治疗→结果的决策路径,辅助临床决策支持系统(CDSS)
+
+```yaml
+type: hypergraph
+核心要素:
+ - patient: 患者特征
+ - diagnosis: 诊断
+ - treatment: 治疗方案
+ - outcome: 治疗结果
+关系:
+ - indicates: 适应症
+ - contraindicated_with: 禁忌组合
+```
+
+---
+
+##### 3.2 `drug_interaction` - 药物相互作用图
+
+**用于文档**:药品说明书、药物相互作用手册、临床药学文档
+
+**建模内容**:药物A、药物B、相互作用类型(协同/拮抗)、作用机制、严重程度
+
+**可视化什么**:展示多种药物之间的相互作用网络,辅助用药安全和药物重整
+
+```yaml
+type: graph
+核心要素:
+ - drug: 药物名称
+ - drug_class: 药物类别
+关系:
+ - interacts_with: 相互作用(标注类型和严重程度)
+ - contraindicated: 禁忌组合
+```
+
+---
+
+##### 3.3 `anatomy_graph` - 解剖结构图
+
+**用于文档**:解剖学教科书、解剖图谱、外科手术记录
+
+**建模内容**:解剖结构名称、结构层级(系统→器官→组织→细胞)、位置
+
+**可视化什么**:展示人体结构的层级包含关系和空间位置,辅助外科计划和医学教育
+
+```yaml
+type: graph
+核心要素:
+ - structure: 解剖结构
+ - level: 层级(系统→器官→组织→细胞)
+ - location: 空间位置
+关系:
+ - contains: 包含关系
+ - located_in: 位于关系
+```
+
+---
+
+##### 3.4 `hospital_timeline` - 住院病程时间线
+
+**用于文档**:出院小结、病程记录、手术记录、会诊记录
+
+**建模内容**:事件类型(入院/检查/手术/用药/出院)、事件日期、执行人、结果
+
+**可视化什么**:展示患者住院期间的事件顺序,便于病历分析和质量控制
+
+```yaml
+type: temporal_graph
+核心要素:
+ - event: 事件(入院/检查/手术/用药/出院)
+ - event_date: 发生日期
+ - performer: 执行人
+关系:
+ - follows: 顺序关系
+ - results_in: 结果关系
+```
+
+---
+
+##### 3.5 `discharge_instruction` - 出院指导
+
+**用于文档**:出院小结、患者教育材料、用药指导单
+
+**建模内容**:用药列表及用法、随访安排、注意事项、警告信号
+
+**可视化什么**:结构化呈现出院后患者需知内容,便于医患沟通和患者管理
+
+```yaml
+type: model
+核心字段:
+ - medications[]: 用药列表及用法
+ - follow_up: 随访安排
+ - restrictions: 注意事项
+ - warning_signs[]: 警告信号
+```
+
+---
+
+### 删除项
+- ❌ `ComplexInteractionNet` → 合并到drug_interaction
+- ❌ `PathologyHypergraph` → 合并到treatment_map
+- ❌ `LevelOfEvidence` → 过于学术
+- ❌ `MicroscopicFeatureSet` → 过于专业
+- ❌ `TumorStagingItem` → 可作为treatment_map子集
+- ❌ 其他10+个模板 → 删除
+
+**medicine预期数量:5个(原15+个)**
+
+---
+
+## ⚖️ 4. **legal** - 法律
+
+### 设计原则
+法律文档逻辑严密,保留核心模板并简化。
+
+### 模板清单(5个)
+
+##### 4.1 `contract_obligation` - 合同义务图
+
+**用于文档**:服务合同(MSA/SLA)、采购合同、租赁合同、劳动合同
+
+**建模内容**:合同主体、义务内容、触发条件、例外情形、违约责任
+
+**可视化什么**:展示合同中各方的权利义务关系、条件触发链,便于合同审查和履约监控
+
+```yaml
+type: hypergraph
+核心要素:
+ - party: 主体
+ - obligation: 义务
+ - trigger: 触发条件
+ - exception: 例外情形
+关系:
+ - must_perform: 主体需履行
+ - triggered_by: 条件触发
+ - exempt_by: 例外免责
+```
+
+---
+
+##### 4.2 `case_fact_timeline` - 案件事实时间线
+
+**用于文档**:判决书、起诉状、答辩状、证据材料
+
+**建模内容**:事实描述、发生日期、相关证据、是否争议、认定结果
+
+**可视化什么**:展示案件事实的时间顺序和证据链,辅助诉讼分析
+
+```yaml
+type: temporal_graph
+核心要素:
+ - fact: 事实描述
+ - date: 发生日期
+ - evidence: 相关证据
+ - disputed: 是否争议
+关系:
+ - preceded_by: 时间先后
+ - proves: 证据支持
+```
+
+---
+
+##### 4.3 `case_citation` - 判例引用图
+
+**用于文档**:判决书、法律意见书、学术论文、判例汇编
+
+**建模内容**:案件名称、引用条款、审理法院、判决结果
+
+**可视化什么**:展示判例之间的引用关系和效力层级,辅助法律研究和先例分析
+
+```yaml
+type: graph
+核心要素:
+ - case: 案件名称
+ - citation: 引用条款
+ - court: 审理法院
+关系:
+ - cites: 引用关系
+ - distinguished_from: 区分解释
+ - overruled_by: 推翻
+```
+
+---
+
+##### 4.4 `compliance_list` - 合规要求清单
+
+**用于文档**:合规手册、法规汇编、审计报告、内控文档
+
+**建模内容**:合规要求、依据法规、完成期限、责任人
+
+**可视化什么**:结构化列出合规要求,便于合规审计和GAP分析
+
+```yaml
+type: list
+核心字段:
+ - requirement: 合规要求
+ - regulation: 依据法规
+ - deadline: 完成期限
+ - responsible_party: 责任人
+```
+
+---
+
+##### 4.5 `defined_term_set` - 定义术语集合
+
+**用于文档**:合同(特别是英文合同)、法律意见书、法规定义章节
+
+**建模内容**:定义术语、定义位置、含义摘要
+
+**可视化什么**:提取合同中所有定义条款,便于术语一致性检查和快速理解
+
+```yaml
+type: set
+核心要素:
+ - term: 定义术语
+ - defined_in: 定义位置
+ - meaning: 含义摘要
+```
+
+---
+
+### 删除项
+- ❌ `AdjudicationLogic` → 过于复杂
+- ❌ `BeneficialOwnershipGraph` → 合并到finance的ownership_graph
+- ❌ `LiabilityClauseList` → 合并到contract_obligation
+- ❌ 其他5+个模板 → 删除
+
+**legal预期数量:5个(原10+个)**
+
+---
+
+## 🌿 5. **tcm** - 中医
+
+### 设计原则
+中医领域小众但专业,选择核心模板并增加关联图。
+
+### 模板清单(5个)
+
+##### 5.1 `formula_composition` - 方剂组成图
+
+**用于文档**:经典方剂著作(《伤寒论》《金匮要略》)、现代方剂学教材、临床验方
+
+**建模内容**:方剂名称、药材名称、角色(君/臣/佐/使)、剂量、配伍关系
+
+**可视化什么**:展示方剂中各药材的角色分配和配伍关系,辅助方剂分析和药理研究
+
+```yaml
+type: hypergraph
+核心要素:
+ - formula: 方剂名称
+ - herb: 药材名称
+ - role: 角色(君/臣/佐/使)
+ - dosage: 剂量
+关系:
+ - combines_with: 配伍关系
+ - antagonizes: 相恶相反
+```
+
+---
+
+##### 5.2 `syndrome_reasoning` - 证候推理图
+
+**用于文档**:医案(医话/医论)、临床经验集、中医诊疗指南
+
+**建模内容**:症状表现、证候类型、治法原则、推荐方剂
+
+**可视化什么**:展示症状→证候→治法→方剂的推理链条,辅助辨证论治和经验挖掘
+
+```yaml
+type: hypergraph
+核心要素:
+ - symptom: 症状表现
+ - syndrome: 证候类型
+ - principle: 治法原则
+ - formula: 推荐方剂
+关系:
+ - leads_to: 导致
+ - treated_by: 治疗
+```
+
+---
+
+##### 5.3 `meridian_graph` - 经络流注图
+
+**用于文档**:针灸学教材、经络腧穴学、针灸临床指南
+
+**建模内容**:穴位名称、所属经脉、位置描述、流注顺序
+
+**可视化什么**:展示穴位在经脉上的分布和流注顺序,辅助针灸教学和临床取穴
+
+```yaml
+type: graph
+核心要素:
+ - acupoint: 穴位名称
+ - meridian: 所属经脉
+ - location: 位置描述
+关系:
+ - located_on: 位于经脉
+ - flows_to: 流注顺序
+```
+
+---
+
+##### 5.4 `herb_property` - 药材属性模型
+
+**用于文档**:中药学教材(《神农本草经》《本草纲目》)、中药数据库、药品标准
+
+**建模内容**:药材名称、四气、五味、归经、功效、毒性
+
+**可视化什么**:结构化呈现药材的基本属性,便于中药数据库建设和临床查询
+
+```yaml
+type: model
+核心字段:
+ - name: 药材名称
+ - nature: 四气(寒/热/温/凉)
+ - flavor: 五味(酸/苦/甘/辛/咸)
+ - meridian_entry: 归经
+ - efficacy: 功效
+ - toxicity: 毒性
+```
+
+---
+
+##### 5.5 `herb_relation` - 药材关联图
+
+**用于文档**:中药学教材、方剂学文献、临床用药指南、药对研究
+
+**建模内容**:药材名称、关联类型(相须/相使/相畏/相杀/相恶/相反)、关联描述
+
+**可视化什么**:展示药材之间的协同或对抗关系,辅助临床配伍和用药安全
+
+```yaml
+type: graph
+核心要素:
+ - herb: 药材名称
+ - herb_property: 药材属性(可选)
+关系:
+ - synergizes: 相须/相使(协同)
+ - reduces_toxicity: 相畏/相杀(减毒)
+ - antagonizes: 相恶/相反(对抗)
+```
+
+---
+
+### 删除项
+- ❌ `AcupointLocationMap` → 合并到meridian_graph
+- ❌ `PulseTongueRecord` → 过于细节
+- ❌ `ProcessingMethod` → 可作为herb_property扩展
+- ❌ 其他5+个模板 → 删除
+
+**tcm预期数量:5个(原10+个)**
+
+---
+
+## ⚙️ 6. **industry** - 工业
+
+### 设计原则
+工业领域文档实用性强,保留操作相关模板。
+
+### 模板清单(5个)
+
+##### 6.1 `operation_flow` - 操作流程图
+
+**用于文档**:SOP操作手册、设备操作规程、维修作业指导书
+
+**建模内容**:操作步骤名称、步骤类型(动作/检查/切换)、设备状态、预期结果
+
+**可视化什么**:展示操作的步骤顺序、状态变化和预期结果,辅助SOP可视化和操作培训
+
+```yaml
+type: graph
+核心要素:
+ - step: 操作步骤名称
+ - step_type: 类型(动作/检查/切换)
+ - equipment_state: 设备状态
+ - expected_result: 预期结果
+关系:
+ - next: 顺序执行
+ - triggers: 触发条件
+ - results_in: 导致状态变化
+```
+
+---
+
+##### 6.2 `safety_control` - 安全控制图
+
+**用于文档**:安全管理手册、安全操作规程、风险评估报告、应急预案
+
+**建模内容**:危险源、风险点、控制措施、责任角色
+
+**可视化什么**:展示危险源→风险→控制措施的链条,辅助安全管理和审核
+
+```yaml
+type: graph
+核心要素:
+ - hazard: 危险源
+ - risk_point: 风险点
+ - control_measure: 控制措施
+ - responsible_role: 责任角色
+关系:
+ - causes: 导致风险
+ - mitigates: 减轻风险
+ - assigned_to: 分配责任
+```
+
+---
+
+##### 6.3 `failure_case` - 故障案例图
+
+**用于文档**:故障分析报告、维修记录、设备履历、故障案例库
+
+**建模内容**:故障现象、根本原因、纠正措施、预防措施、经验教训
+
+**可视化什么**:展示故障现象→原因→措施的因果链,辅助故障诊断和经验积累
+
+```yaml
+type: graph
+核心要素:
+ - phenomenon: 故障现象
+ - root_cause: 根本原因
+ - corrective_action: 纠正措施
+ - lesson: 经验教训
+关系:
+ - caused_by: 由...引起
+ - resolved_by: 通过...解决
+ - prevented_by: 预防措施
+```
+
+---
+
+##### 6.4 `equipment_topology` - 设备拓扑图
+
+**用于文档**:设备手册、系统设计文档、工厂布局图、备件清单
+
+**建模内容**:设备名称、设备类型、安装位置、规格参数、连接关系
+
+**可视化什么**:展示设备之间的物理连接和层级关系,辅助设备管理和系统理解
+
+```yaml
+type: graph
+核心要素:
+ - equipment: 设备名称
+ - equipment_type: 设备类型
+ - location: 安装位置
+ - specifications: 规格参数
+关系:
+ - connected_to: 物理连接
+ - located_in: 安装位置
+ - contains: 包含关系
+```
+
+---
+
+##### 6.5 `emergency_response` - 应急响应图
+
+**用于文档**:应急预案、应急演练方案、事故处置规程
+
+**建模内容**:事故场景、响应动作、责任人、响应顺序、资源需求
+
+**可视化什么**:展示事故→响应动作→责任人的触发链,辅助应急演练和事故处置
+
+```yaml
+type: graph
+核心要素:
+ - scenario: 事故场景
+ - action: 响应动作
+ - responsible_role: 责任人
+ - resource: 所需资源
+关系:
+ - triggers: 触发响应
+ - executed_by: 执行责任人
+ - requires: 所需资源
+```
+
+---
+
+### 删除项
+- ❌ `IncidentCausalityMap` → 合并到failure_case
+- ❌ `SystemCompatibilityGraph` → 使用率低
+- ❌ `PartReplacementList` → 过于细节
+- ❌ `SystemTopologyGraph` → 与equipment_topology重复
+- ❌ 其他10+个模板 → 删除
+
+**industry预期数量:5个(原15+个)**
+
+---
+
+## 📊 重构预期效果
+
+### 模板数量变化
+
+| 领域 | 重构前 | 重构后 | 变化 |
+|------|--------|--------|------|
+| general | 14 | 13 | +4新增,-5删除 |
+| finance | 20+ | 5 | -75% |
+| medicine | 15+ | 5 | -67% |
+| legal | 10+ | 5 | -50% |
+| tcm | 10+ | 5 | -50% |
+| industry | 15+ | 5 | -67% |
+| **总计** | **80+** | **38** | **-53%** |
+
+### 新增/重命名模板汇总
+
+| 新模板 | 用途 | 来源 |
+|--------|------|------|
+| `workflow_graph` | 流程图(Skill/Workflow/SOP) | 新增 |
+| `doc_structure` | 文档结构图 | 新增 |
+| `biography_graph` | 传记图 | 重命名(life_event_timeline) |
+| `concept_graph` | 概念关系图 | 重命名(concept_hierarchy) |
+| `herb_relation` | 药材关联图 | 新增 |
+
+---
+
+## 🗓️ 实施计划
+
+### Phase 1: 准备阶段
+- [ ] 确认保留的base_*模板清单
+- [ ] 制定各模板YAML格式规范
+- [ ] 建立模板测试用例
+
+### Phase 2: general领域(优先)
+- [ ] workflow_graph
+- [ ] doc_structure
+- [ ] biography_graph
+- [ ] concept_graph
+
+### Phase 3: 其他领域精简
+- [ ] finance:保留5个,删除15+
+- [ ] medicine:保留5个,删除10+
+- [ ] legal:保留5个,删除5+
+- [ ] tcm:保留5个,新增herb_relation,删除5+
+- [ ] industry:保留5个,删除10+
+
+### Phase 4: legacy清理
+- [ ] 将归档领域移至legacy/presets
+- [ ] 清理Python模板(zh目录)
+- [ ] 更新文档引用
+
+### Phase 5: 验证与发布
+- [ ] 运行模板测试
+- [ ] 更新README
+- [ ] 发布新版本
+
+---
+
+## 🤔 待确认事项
+
+1. **workflow_graph的粒度**:
+ - 是否需要支持嵌套的subskill/step?
+ - 条件分支的表示方式?(if/else/loop)
+ - 是否需要支持并行执行节点?
+
+2. **biography_graph**:
+ - 是否需要支持"人物关系"作为单独的关系类型?
+ - 地点信息用spatial_graph还是作为event的属性?
+
+3. **herb_relation关联类型**:
+ - 是否需要区分"相须/相使/相畏/相杀/相恶/相反"?
+
+4. **模板优先级**:哪些模板应该优先实现?
+
+5. **legacy清理时机**:是现在清理还是分阶段?
+
+---
+
+## 📝 附录:GitHub热点参考
+
+### 2026年3月Trending项目
+| 项目 | Stars | 领域 |
+|------|-------|------|
+| DeerFlow | 47K | AI Agent |
+| OpenClaw | 270K | AI Assistant |
+| ruflo | 26K | Agent Orchestration |
+| obra/superpowers | 15K | AI Coding OS |
diff --git a/hyperextract/templates/presets/finance/.gitkeep b/hyperextract/templates/presets/finance/.gitkeep
deleted file mode 100644
index e69de29..0000000
diff --git a/hyperextract/templates/presets/finance/earnings_summary.yaml b/hyperextract/templates/presets/finance/earnings_summary.yaml
new file mode 100644
index 0000000..e650ae9
--- /dev/null
+++ b/hyperextract/templates/presets/finance/earnings_summary.yaml
@@ -0,0 +1,104 @@
+language: [zh, en]
+
+name: earnings_summary
+type: model
+tags: [finance, earnings, investor-relations]
+
+description:
+ zh: '财报会议摘要模板 - 从财报电话会议记录中提取公司业绩、预期对比和前瞻指引'
+ en: 'Earnings Call Summary Template - Extract company performance, consensus comparison, and guidance from earnings call transcripts'
+
+output:
+ description:
+ zh: '财报会议的核心内容和关键信息'
+ en: 'Core content and key information from earnings calls'
+ fields:
+ - name: company_name
+ type: str
+ description:
+ zh: '公司名称'
+ en: 'Company name'
+ required: true
+ - name: quarter
+ type: str
+ description:
+ zh: '财报季度,格式如"2024Q3"'
+ en: 'Reporting quarter in format like "2024Q3"'
+ required: true
+ - name: reported_revenue
+ type: str
+ description:
+ zh: '报告期实际营收'
+ en: 'Reported revenue for the period'
+ required: true
+ - name: reported_eps
+ type: str
+ description:
+ zh: '报告期实际每股收益'
+ en: 'Reported earnings per share'
+ required: true
+ - name: revenue_vs_consensus
+ type: str
+ description:
+ zh: '营收与市场预期的对比'
+ en: 'Revenue comparison with market consensus'
+ required: true
+ - name: eps_vs_consensus
+ type: str
+ description:
+ zh: '每股收益与市场预期的对比'
+ en: 'EPS comparison with market consensus'
+ required: true
+ - name: guidance_revenue
+ type: str
+ description:
+ zh: '营收前瞻指引'
+ en: 'Revenue guidance'
+ required: false
+ - name: guidance_eps
+ type: str
+ description:
+ zh: '每股收益前瞻指引'
+ en: 'EPS guidance'
+ required: false
+ - name: overall_tone
+ type: str
+ description:
+ zh: '管理层整体语调:积极/中性/消极'
+ en: 'Overall management tone: positive/neutral/negative'
+ required: true
+ - name: key_highlights
+ type: list
+ description:
+ zh: '本次会议的主要亮点'
+ en: 'Key highlights from this call'
+ required: false
+ - name: key_concerns
+ type: list
+ description:
+ zh: '本次会议的主要关切问题'
+ en: 'Key concerns raised in this call'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的投研分析师,负责从财报电话会议记录中准确提取业绩数据和关键信息。'
+ en: 'You are a professional investment research analyst responsible for accurately extracting performance data and key information from earnings call transcripts.'
+ rules:
+ zh:
+ - '提取精确的营收和EPS数字,确保与财报一致'
+ - '对比数据应反映与市场共识预期的差异'
+ - '前瞻指引应区分营收指引和EPS指引'
+ - '语调评估应基于管理层措辞和信心程度'
+ - '亮点和关切应提炼自会议讨论的核心议题'
+ en:
+ - 'Extract precise revenue and EPS figures, ensuring consistency with financial reports'
+ - 'Comparison data should reflect the difference from market consensus expectations'
+ - 'Guidance should distinguish between revenue and EPS guidance'
+ - 'Tone assessment should be based on management language and confidence level'
+ - 'Highlights and concerns should distill the core topics discussed in the call'
+
+identifiers: {}
+
+display:
+ label: '{company_name} {quarter}'
diff --git a/hyperextract/templates/presets/finance/event_timeline.yaml b/hyperextract/templates/presets/finance/event_timeline.yaml
new file mode 100644
index 0000000..4ec0336
--- /dev/null
+++ b/hyperextract/templates/presets/finance/event_timeline.yaml
@@ -0,0 +1,105 @@
+language: [zh, en]
+
+name: event_timeline
+type: temporal_graph
+tags: [finance, event, timeline, corporate-events]
+
+description:
+ zh: '重大事件时间线模板 - 从8-K报告、新闻稿和公司公告中提取公司重大事件的时间线'
+ en: 'Major Events Timeline Template - Extract corporate major event timelines from 8-K reports, press releases, and company announcements'
+
+output:
+ description:
+ zh: '公司重大事件的时间顺序和因果关系'
+ en: 'Temporal order and causal relationships of corporate major events'
+ entities:
+ description:
+ zh: '公司及其相关事件'
+ en: 'Company and related events'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '公司或事件名称'
+ en: 'Company or event name'
+ required: true
+ - name: entity_type
+ type: str
+ description:
+ zh: '实体类型:公司/事件/公告类型'
+ en: 'Entity type: company/event/announcement type'
+ required: true
+ relations:
+ description:
+ zh: '事件之间的时间顺序和因果关系'
+ en: 'Temporal sequence and causal relationships between events'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '源实体名称'
+ en: 'Source entity name'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '目标实体名称'
+ en: 'Target entity name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:precedes/causes/triggers'
+ en: 'Relation type: precedes/causes/triggers'
+ required: true
+ - name: time
+ type: str
+ description:
+ zh: '事件日期'
+ en: 'Event date'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '事件说明,包括是否构成重大事项'
+ en: 'Event description, including whether it constitutes a material event'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的企业事件分析师,负责从公司披露文件中提取重大事件并构建时间线。'
+ en: 'You are a professional corporate events analyst responsible for extracting major events and constructing timelines from company disclosure documents.'
+ rules_for_entities:
+ zh:
+ - '提取公司名称作为核心实体'
+ - '提取具有重大意义的事件作为实体'
+ - '事件类型应反映事件的性质(公告/评级调整/业绩发布等)'
+ en:
+ - 'Extract company name as the core entity'
+ - 'Extract events of significant meaning as entities'
+ - 'Event type should reflect the nature of the event (announcement/rating adjustment/earnings release, etc.)'
+ rules_for_relations:
+ zh:
+ - 'type为precedes表示事件的时间先后顺序'
+ - 'type为causes表示因果关联'
+ - 'type为triggers表示触发关系'
+ - '仅提取文本中明确说明的关系'
+ - 'description中应说明事件是否构成重大事项'
+ en:
+ - 'type precedes indicates temporal sequence of events'
+ - 'type causes indicates causal connection'
+ - 'type triggers indicates triggering relationship'
+ - 'Only extract explicitly stated relationships in the text'
+ - 'Description should indicate whether the event constitutes a material event'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}|{time}'
+ relation_members:
+ source: source
+ target: target
+ time_field: time
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/finance/ownership_graph.yaml b/hyperextract/templates/presets/finance/ownership_graph.yaml
new file mode 100644
index 0000000..82514d9
--- /dev/null
+++ b/hyperextract/templates/presets/finance/ownership_graph.yaml
@@ -0,0 +1,102 @@
+language: [zh, en]
+
+name: ownership_graph
+type: graph
+tags: [finance, corporate, ownership, shareholder]
+
+description:
+ zh: '股权结构图模板 - 从招股说明书和年报中提取股东持股关系和控制结构'
+ en: 'Ownership Structure Graph Template - Extract shareholder ownership relationships and control structure from IPO prospectuses and annual reports'
+
+output:
+ description:
+ zh: '公司股权层级结构和管理关系'
+ en: 'Corporate ownership hierarchy and control relationships'
+ entities:
+ description:
+ zh: '股东和公司实体'
+ en: 'Shareholder and company entities'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '股东或公司名称'
+ en: 'Shareholder or company name'
+ required: true
+ - name: shareholder_type
+ type: str
+ description:
+ zh: '股东类型:机构投资者/个人/关联方/上市公司'
+ en: 'Shareholder type: institutional investor/individual/related party/public company'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '股东或公司说明'
+ en: 'Shareholder or company description'
+ required: false
+ relations:
+ description:
+ zh: '股东持股和控制关系'
+ en: 'Shareholding and control relationships'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '控股股东名称'
+ en: 'Controlling shareholder name'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '被控股公司名称'
+ en: 'Controlled company name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:controls/belongs_to/holds_shares'
+ en: 'Relation type: controls/belongs_to/holds_shares'
+ required: true
+ - name: ownership_percentage
+ type: str
+ description:
+ zh: '持股比例'
+ en: 'Ownership percentage'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的企业分析师,负责从IPO招股说明书和年报中准确提取股权结构信息。'
+ en: 'You are a professional corporate analyst responsible for accurately extracting ownership structure information from IPO prospectuses and annual reports.'
+ rules_for_entities:
+ zh:
+ - '提取具有独立法律地位的股东实体'
+ - '股东类型应根据实际性质分类'
+ - '机构投资者应标注具体类型(基金/保险/券商等)'
+ en:
+ - 'Extract shareholder entities with independent legal status'
+ - 'Shareholder type should be classified based on actual nature'
+ - 'Institutional investors should specify type (fund/insurance/securities, etc.)'
+ rules_for_relations:
+ zh:
+ - '优先提取文本中明确描述的持股和控制关系'
+ - '持股比例应与原文保持一致'
+ - '注意区分直接持股和间接持股'
+ - '仅提取文本中明确表达的关系'
+ en:
+ - 'Prioritize explicitly described shareholding and control relationships in the text'
+ - 'Ownership percentage should match the source text'
+ - 'Distinguish between direct and indirect shareholding'
+ - 'Only extract relationships explicitly expressed in the text'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type} {ownership_percentage}'
diff --git a/hyperextract/templates/presets/finance/risk_factor_set.yaml b/hyperextract/templates/presets/finance/risk_factor_set.yaml
new file mode 100644
index 0000000..095405f
--- /dev/null
+++ b/hyperextract/templates/presets/finance/risk_factor_set.yaml
@@ -0,0 +1,55 @@
+language: [zh, en]
+
+name: risk_factor_set
+type: set
+tags: [finance, risk, risk-management]
+
+description:
+ zh: '风险因子集合模板 - 从年报风险章节中提取和归类风险因子'
+ en: 'Risk Factor Set Template - Extract and categorize risk factors from annual report risk sections'
+
+output:
+ description:
+ zh: '归类后的风险因子清单'
+ en: 'Categorized risk factor list'
+ fields:
+ - name: risk_factor
+ type: str
+ description:
+ zh: '风险因子描述'
+ en: 'Risk factor description'
+ required: true
+ - name: category
+ type: str
+ description:
+ zh: '风险类别:经营风险/财务风险/市场风险/法规风险'
+ en: 'Risk category: operational/financial/market/regulatory'
+ required: true
+ - name: severity
+ type: str
+ description:
+ zh: '严重程度:高/中/低'
+ en: 'Severity: high/medium/low'
+ required: true
+
+guideline:
+ target:
+ zh: '你是一位专业的风险分析师,负责从年报风险章节中准确提取和归类风险因子。'
+ en: 'You are a professional risk analyst responsible for accurately extracting and categorizing risk factors from annual report risk sections.'
+ rules:
+ zh:
+ - '提取的风险因子应完整保留原文的核心含义'
+ - '风险类别应根据因子性质准确归类'
+ - '严重程度应基于风险发生的可能性和影响程度评估'
+ - '同一类型的风险因子应保持归类一致性'
+ en:
+ - 'Extracted risk factors should fully preserve the core meaning from the source text'
+ - 'Risk category should be accurately classified based on the nature of the factor'
+ - 'Severity should be assessed based on likelihood and impact of risk occurrence'
+ - 'Risk factors of the same type should maintain classification consistency'
+
+identifiers:
+ item_id: risk_factor
+
+display:
+ label: '{risk_factor}'
diff --git a/hyperextract/templates/presets/finance/sentiment_model.yaml b/hyperextract/templates/presets/finance/sentiment_model.yaml
new file mode 100644
index 0000000..d0afa9d
--- /dev/null
+++ b/hyperextract/templates/presets/finance/sentiment_model.yaml
@@ -0,0 +1,62 @@
+language: [zh, en]
+
+name: sentiment_model
+type: model
+tags: [finance, sentiment, market-analysis]
+
+description:
+ zh: '市场情绪模型模板 - 从财经新闻和市场评论中提取量化情绪指标'
+ en: 'Market Sentiment Model Template - Extract quantified sentiment indicators from financial news and market commentary'
+
+output:
+ description:
+ zh: '市场情绪的量化评估结果'
+ en: 'Quantified market sentiment assessment result'
+ fields:
+ - name: sentiment_score
+ type: float
+ description:
+ zh: '情绪分数,范围从-1(极度负面)到1(极度正面)'
+ en: 'Sentiment score ranging from -1 (extremely negative) to 1 (extremely positive)'
+ required: true
+ - name: dominant_theme
+ type: str
+ description:
+ zh: '主导市场主题'
+ en: 'Dominant market theme'
+ required: true
+ - name: key_entities
+ type: list
+ description:
+ zh: '文本中提到的关键实体(公司、行业、产品等)'
+ en: 'Key entities mentioned in the text (companies, industries, products, etc.)'
+ required: false
+ - name: price_impact
+ type: str
+ description:
+ zh: '对价格/市场的影响预测'
+ en: 'Predicted impact on price/market'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的市场情绪分析师,负责从财经文本中量化评估市场情绪。'
+ en: 'You are a professional market sentiment analyst responsible for quantifying market sentiment from financial texts.'
+ rules:
+ zh:
+ - '情绪分数应客观反映文本整体情绪倾向'
+ - '-1到-0.3为负面,-0.3到0.3为中性,0.3到1为正面'
+ - '主导主题应提炼文本的核心议题'
+ - '关键实体应包括所有重要的公司和行业提及'
+ - '价格影响应基于文本对市场的预期判断'
+ en:
+ - 'Sentiment score should objectively reflect the overall emotional tendency of the text'
+ - '-1 to -0.3 is negative, -0.3 to 0.3 is neutral, 0.3 to 1 is positive'
+ - 'Dominant theme should distill the core topic of the text'
+ - 'Key entities should include all important company and industry mentions'
+ - 'Price impact should be based on the text expectations for the market'
+
+identifiers: {}
+
+display:
+ label: '{dominant_theme}'
diff --git a/hyperextract/templates/presets/general/biography_graph.yaml b/hyperextract/templates/presets/general/biography_graph.yaml
new file mode 100644
index 0000000..68adc6d
--- /dev/null
+++ b/hyperextract/templates/presets/general/biography_graph.yaml
@@ -0,0 +1,110 @@
+language: [zh, en]
+
+name: biography_graph
+type: temporal_graph
+tags: [general, biography, life, events, timeline]
+
+description:
+ zh: '传记图模板 - 从人物传记、回忆录、年谱中提取实体和关系'
+ en: 'Biography Graph Template - Extract entities and relationships from biographies, memoirs, and chronicles'
+
+output:
+ description:
+ zh: '人物传记中的实体和关系网络'
+ en: "Entity and relationship network from biographies"
+ entities:
+ description:
+ zh: '传记中的实体'
+ en: 'Entities in the biography'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '实体名称'
+ en: 'Entity name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '实体类型,如人物、地点、组织、作品、事件、成就等'
+ en: 'Entity type, such as person, location, organization, work, event, achievement, etc.'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '实体描述'
+ en: 'Entity description'
+ required: false
+ relations:
+ description:
+ zh: '实体之间的关系'
+ en: 'Relationships between entities'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '关系起点实体'
+ en: 'Source entity'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '关系终点实体'
+ en: 'Target entity'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型,用简洁词语概括两个实体之间的连接方式'
+ en: 'Relation type, concisely describing the connection between two entities'
+ required: true
+ - name: time
+ type: str
+ description:
+ zh: '关系对应的时间;无明确时间时可留空'
+ en: 'Time associated with the relation; can be left empty if no explicit time'
+ required: false
+ - name: description
+ type: str
+ description:
+ zh: '对关系依据、方向或上下文含义的简要说明'
+ en: 'Brief description of relation basis, direction, or contextual meaning'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的传记分析师,擅长从人物传记、回忆录和年谱中提取实体和关系。'
+ en: 'You are a professional biography analyst skilled at extracting entities and relationships from biographies, memoirs, and chronicles.'
+ rules_for_entities:
+ zh:
+ - '提取传记主人公作为核心实体'
+ - '相关人物包括:家人、朋友、同事、对手、历史人物等'
+ - '同一实体在全文中保持命名一致'
+ - '仅提取文本中明确表达的实体'
+ en:
+ - 'Extract the biographical subject as the core entity'
+ - 'Related persons include: family, friends, colleagues, rivals, historical figures'
+ - 'Maintain consistent naming for the same entity throughout the text'
+ - 'Only extract entities explicitly expressed in the text'
+ rules_for_relations:
+ zh:
+ - '提取对理解人物生平有重要意义的关系'
+ - '仅在文本明确表达两个实体之间存在关系时创建关系'
+ - '优先使用文本中明确出现的关系词'
+ - '不要创建隐含的常识性关联'
+ en:
+ - 'Extract relationships of significant importance for understanding the person life'
+ - 'Create relationships only when the text explicitly expresses a relationship between two entities'
+ - 'Prefer relation words that explicitly appear in the text'
+ - 'Do not create implicit common-sense associations'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/general/clause_list.yaml b/hyperextract/templates/presets/general/clause_list.yaml
deleted file mode 100644
index d78f9bb..0000000
--- a/hyperextract/templates/presets/general/clause_list.yaml
+++ /dev/null
@@ -1,50 +0,0 @@
-language: zh
-name: clause_list
-type: list
-tags: [general]
-description:
- zh: '规章条文清单 - 将规章制度、操作手册或合规文本拆解为逐条的原子化条文,便于索引、检索、比对与问答。'
-
-output:
- description:
- zh: '由逐条拆解后的规章条文组成的列表结果。'
- fields:
- - name: clauseId
- type: str
- description:
- zh: '条文编号,如第1条、第2条、第2.1款、第3项等。'
- - name: title
- type: str
- description:
- zh: '条文标题、主题或该条的简短概括;原文未明确给出时可留空。'
- required: false
- default: ''
- - name: content
- type: str
- description:
- zh: '条文正文内容,尽量保持与原文一致,避免改写。'
- - name: category
- type: str
- description:
- zh: '条文分类,如总则、权利义务、程序、处罚、附则、其他。'
- required: false
- default: '其他'
-
-guideline:
- target:
- zh: '你是一位专业的规章制度条文拆解专家。请将输入文本按其自然结构拆解为可独立引用、可检索的逐条条文。'
- rules:
- zh:
- - '按原文的自然层级和编号结构拆解条文,如第X条、第X款、第X项等。'
- - '每一项输出都应对应一个独立、可单独引用的条文单元。'
- - 'clauseId 必须尽量使用原文中的编号;若原文无明确编号,可使用稳定且简洁的顺序标识。'
- - 'content 应保留该条的核心正文,尽量与原文一致,不要总结式改写。'
- - 'title 仅在原文明确存在标题、主题或能够从该条主旨中直接提炼时填写;否则留空。'
- - 'category 需尽量归入总则、权利义务、程序、处罚、附则、其他之一。'
- - '不要遗漏任何实质性条文,也不要把多个独立条文合并成一条。'
- - '目录、页眉页脚、签章、编号噪音、纯格式性重复内容通常不作为条文输出。'
-
-options: {}
-
-display:
- label: '{clauseId}({title})'
diff --git a/hyperextract/templates/presets/general/compliance_logic.yaml b/hyperextract/templates/presets/general/compliance_logic.yaml
deleted file mode 100644
index eb66250..0000000
--- a/hyperextract/templates/presets/general/compliance_logic.yaml
+++ /dev/null
@@ -1,117 +0,0 @@
-language: zh
-name: compliance_logic
-type: hypergraph
-tags: [general]
-description:
- zh: '规定条文超图 - 面向制度、细则、办法、规范、指引、规程等规定类文本,提取条文中涉及的名词性要素,并用 action 概括这些要素之间被要求执行的行为。'
-
-output:
- description:
- zh: '由名词性实体节点和规定超边组成的结构化结果。节点只保留规定中涉及的人、部门、文件、场所、物品等要素;超边表示一条独立规定,action 用于概括这些要素之间被要求执行的行为。'
- entities:
- description:
- zh: '规定涉及的基础节点。只提取名词或名词短语,不提取动作本身。'
- fields:
- - name: name
- type: str
- description:
- zh: '实体名称,尽量直接使用原文中的稳定称呼;同一对象若同时出现全称、简称或缩写,应统一为一个名称。'
- - name: category
- type: str
- description:
- zh: '实体类型,应尽量使用清晰、稳定的名词类别,如人员、部门、组织、岗位、文件表单、证件、场所、设备、物品、时间要求、其他。'
- - name: description
- type: str
- description:
- zh: '用一句话说明该实体在当前规定中的含义或作用。'
- required: false
- default: ''
- relations:
- description:
- zh: '表达规定条文的超边。每条超边对应一条独立规定,由相关实体构成,action 负责概括这些实体之间被要求执行的行为。'
- fields:
- - name: ruleType
- type: str
- description:
- zh: '规定类型,用于概括该条规定在规定什么。优先使用:职责分工、行为要求、流程程序、其他。'
- - name: participants
- type: list
- description:
- zh: '构成该条规定基本内容的实体名称列表,必须直接使用已提取节点的 name。只放实体,不要把动作短语放进 participants。'
- - name: action
- type: str
- description:
- zh: '该条规定涉及的核心行为概括,用 1 句话说明这些实体之间被要求执行什么行为。action 承担行为表达,不要把 action 再拆成实体。'
- - name: sourceClause
- type: str
- description:
- zh: '规定来源的条款编号、标题或出处标识,尽量写最近的编号或小节定位。'
- required: false
- default: ''
-
-guideline:
- target:
- zh: '你是一位规定条文结构化分析专家。请从规定类文本中提取每条规定涉及的名词性要素,并用 action 写清这些要素之间被要求执行的行为。'
- rules_for_entities:
- zh:
- - '提取名词或名词短语形式的实体'
- - '优先提取规定中真正参与行为的要素,如人员、岗位、部门、组织、文件表单、证件、场所、设备、物品、时间要求等。根据文本的具体内容和你的知识来决定具体的类别。'
- - 'name 尽量保持原文中的稳定称呼;同一对象不要重复命名。'
- - '如果某个词只是行为本身,例如“登记”“审批”“填写交接记录”,不要单独作为实体。'
- - '对于“访客证”“登记表”“样品库”“前台”“每日18:00前”这类名词性短语,可以作为实体提取。'
- - 'category 要尽量使用少数、容易理解的类别,不要为了细分而临时创造模糊类别。'
- rules_for_relations:
- zh:
- - '每条超边只表达一条独立规定,并连接这条规定涉及的相关实体。'
- - 'ruleType 优先使用:职责分工、行为要求、流程程序、其他。根据文本的具体内容和你的知识来决定具体的类别。'
- - 'participants 必须直接引用已提取节点的 name,只放名词性实体,不放动作短语。'
- - 'action 用 1 句话概括这条规定要求这些实体之间发生什么行为。'
- - |
- 完整示例(输入):
- 1. 访客进入样品库前,应在前台登记并领取访客证。
- 2. 安全部门每日18:00前回收当日访客证,并存档登记表。
- - |
- 完整示例(输出):
- entities:
- - name: 访客
- category: 人员
- description: 进入样品库并需履行登记要求的外来人员。
- - name: 样品库
- category: 场所
- description: 访客进入前需要满足规定要求的区域。
- - name: 前台
- category: 场所
- description: 办理登记和领取访客证的地点。
- - name: 访客证
- category: 证件
- description: 访客登记后领取并在离开前回收的证件。
- - name: 安全部门
- category: 部门
- description: 负责回收访客证并保存记录的管理部门。
- - name: 登记表
- category: 文件表单
- description: 记录访客登记信息并需要存档的表单。
- - name: 每日18:00前
- category: 时间要求
- description: 安全部门执行回收和存档要求的截止时间要求。
- relations:
- - ruleType: 流程程序
- participants: [访客, 样品库, 前台, 访客证]
- action: 规定访客进入样品库前,应在前台登记并领取访客证。
- sourceClause: 1
- - ruleType: 流程程序
- participants: [安全部门, 访客证, 登记表, 每日18:00前]
- action: 规定安全部门每日18:00前回收当日访客证,并存档登记表。
- sourceClause: 2
-
-identifiers:
- entity_id: name
- relation_id: '{sourceClause}|{ruleType}|{action}'
- relation_members: participants
-
-options:
- extraction_mode: one_stage
-
-display:
- entity_label: '{name} ({category})'
- relation_label: '[{ruleType}] {action}'
diff --git a/hyperextract/templates/presets/general/concept_graph.yaml b/hyperextract/templates/presets/general/concept_graph.yaml
new file mode 100644
index 0000000..2793b6c
--- /dev/null
+++ b/hyperextract/templates/presets/general/concept_graph.yaml
@@ -0,0 +1,102 @@
+language: [zh, en]
+
+name: concept_graph
+type: graph
+tags: [general, concept, knowledge, taxonomy]
+
+description:
+ zh: '概念关系图模板 - 从教科书、百科全书、学术论文中提取概念层级关系和语义关联'
+ en: 'Concept Graph Template - Extract concept hierarchy and semantic associations from textbooks, encyclopedias, and academic papers'
+
+output:
+ description:
+ zh: '概念之间的层级关系和语义关联网络'
+ en: 'Concept hierarchy and semantic association network'
+ entities:
+ description:
+ zh: '文档中的核心概念'
+ en: 'Core concepts in the document'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '概念名称'
+ en: 'Concept name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '概念类型:实体/抽象概念/过程/关系'
+ en: 'Concept type: entity/abstract/process/relation'
+ required: true
+ - name: definition
+ type: str
+ description:
+ zh: '概念的简要定义'
+ en: 'Brief concept definition'
+ required: false
+ relations:
+ description:
+ zh: '概念之间的层级关系和语义关联'
+ en: 'Hierarchy and semantic associations between concepts'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '源概念名称'
+ en: 'Source concept name'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '目标概念名称'
+ en: 'Target concept name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型,用简洁词语概括两个概念之间的连接方式'
+ en: 'Relation type, concisely describing the connection between two concepts'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '关系说明'
+ en: 'Relation description'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位知识图谱专家,擅长从教科书、百科全书和学术论文中提取概念之间的层级关系和语义关联。'
+ en: 'You are a knowledge graph expert skilled at extracting concept hierarchy and semantic associations from textbooks, encyclopedias, and academic papers.'
+ rules_for_entities:
+ zh:
+ - '提取文档中定义的核心概念'
+ - '概念名称应使用规范术语'
+ - '同一概念在全文中保持命名一致'
+ - '仅提取文本中明确表达的概念'
+ en:
+ - 'Extract core concepts defined in the document'
+ - 'Concept names should use standard terminology'
+ - 'Maintain consistent naming for the same concept throughout the text'
+ - 'Only extract concepts explicitly expressed in the text'
+ rules_for_relations:
+ zh:
+ - 'type用简洁词语概括关系性质'
+ - '优先提取文档中明确定义的关系'
+ - '不要创建隐含的常识性关联,仅提取文档中明确说明的关系'
+ en:
+ - 'Type should concisely describe the nature of the relationship'
+ - 'Prioritize explicitly defined relationships in the document'
+ - 'Do not create implicit common-sense associations, only extract explicitly stated relationships'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/general/concept_hierarchy.yaml b/hyperextract/templates/presets/general/concept_hierarchy.yaml
deleted file mode 100644
index 5d6e194..0000000
--- a/hyperextract/templates/presets/general/concept_hierarchy.yaml
+++ /dev/null
@@ -1,78 +0,0 @@
-language: zh
-name: concept_hierarchy
-type: graph
-tags: [general]
-description:
- zh: '概念层级图谱模板 - 从教材、百科、定义说明等文本中提取概念及其层次结果的组成或实例化关系,用于构建稳定的概念结构。'
-
-output:
- description:
- zh: '由概念节点和层级关系边组成的结构化结果,用于表达概念体系中的分类、组成与实例关系。'
- entities:
- description:
- zh: '文本中出现并可进入概念结构的节点。'
- fields:
- - name: name
- type: str
- description:
- zh: '概念名称、术语名称或实例名称,优先使用原文中的稳定叫法。'
- - name: category
- type: str
- description:
- zh: '概念所属专业类别,需要通过文中内容归纳并结合你的常识进行判断'
- - name: description
- type: str
- description:
- zh: '对该概念的简要说明,可概括其定义、范围或在当前结构中的作用。'
- required: false
- default: ''
- relations:
- description:
- zh: '概念之间的层级或结构关系边。'
- fields:
- - name: source
- type: str
- description:
- zh: '关系起点,通常为上位概念、整体概念或被实例化的概念。'
- - name: target
- type: str
- description:
- zh: '关系终点,通常为下位概念、组成部分或实例。'
- - name: relation_type
- type: str
- description:
- zh: '层次关系类型,使用具体层级关系词'
- - name: details
- type: str
- description:
- zh: '对关系方向、限定条件或上下文的简要补充。'
- required: false
- default: ''
-
-guideline:
- target:
- zh: '你是一位概念层级抽取专家。请只提取能够组成层级结构的概念,并同时给出这些概念之间明确成立的层次结构组成或实例关系。'
- rules_for_entities:
- zh:
- - '必须使用原文中的完整稳定名称;不要把"推理/预测"缩写成"推理"这类更短名字。'
- - '同一概念在全文中保持命名一致。'
- - '如果某个概念在全文中找不到任何可靠关系,不要为了保留它而强行构造边。'
- rules_for_relations:
- zh:
- - '只在文本明确表达层级结构时建立边。'
- - '优先保证图谱连通性和关系准确性,而不是追求节点数量。'
- - '关系词应使用文本中明确出现的层次关系词语(如"包含"、"组成"、"是...的实例"等)。'
-
-identifiers:
- entity_id: name
- relation_id: '{source}|{relation_type}|{target}'
- relation_members:
- source: source
- target: target
-
-options:
- extraction_mode: one_stage
-
-display:
- entity_label: '{name} ({category})'
- relation_label: '{relation_type}'
diff --git a/hyperextract/templates/presets/general/cross_reference_net.yaml b/hyperextract/templates/presets/general/cross_reference_net.yaml
deleted file mode 100644
index 42b9da1..0000000
--- a/hyperextract/templates/presets/general/cross_reference_net.yaml
+++ /dev/null
@@ -1,81 +0,0 @@
-language: zh
-name: cross_reference_net
-type: graph
-tags: [general]
-description:
- zh: '交叉引用网络模板 - 从百科、文档、规章或知识说明类文本中提取条目之间的引用、参见、链接或相关关系。'
-
-output:
- description:
- zh: '由条目节点和交叉引用关系组成的网络结构,用于表达文本中的引用路径和知识跳转关系。'
- entities:
- description:
- zh: '可作为引用发起方或引用目标的条目、概念或章节节点。'
- fields:
- - name: name
- type: str
- description:
- zh: '条目、概念、章节或被引用对象的名称。'
- - name: type
- type: str
- description:
- zh: '节点类型,如条目、章节、概念、人物、地点、其他。'
- - name: description
- type: str
- description:
- zh: '对节点含义或其在当前引用网络中角色的简要说明。'
- required: false
- default: ''
- relations:
- description:
- zh: '节点之间的交叉引用关系边。'
- fields:
- - name: source
- type: str
- description:
- zh: '引用发起方,即当前条目、章节或概念。'
- - name: target
- type: str
- description:
- zh: '被引用、被链接、被参见或被关联的目标节点。'
- - name: relation_type
- type: str
- description:
- zh: '关系类型,优先使用超链接、引用、参见、相关、其他。'
- - name: context
- type: str
- description:
- zh: '对引用语境、引用目的或上下文位置的简要补充。'
- required: false
- default: ''
-
-guideline:
- target:
- zh: '你是一位知识网络分析专家。请从文本中提取可独立识别的条目或概念,并建立它们之间明确存在的引用、参见或相关连接。'
- rules_for_entities:
- zh:
- - '优先提取在文本中被明确提及、并能够作为引用源或引用目标的条目、概念、章节、人物或地点。'
- - 'name 使用原文中最清晰、最稳定的称呼;同一对象保持统一命名。'
- - 'type 需概括节点身份,如条目、章节、概念、人物、地点、其他。'
- - 'description 可用于说明该节点为何会进入当前引用网络;信息不足时可留空。'
- rules_for_relations:
- zh:
- - '仅在文本明确存在引用、超链接、参见、相关提示或互指关系时建立边。'
- - 'source 和 target 必须直接引用已提取节点的 name。'
- - 'relation_type 优先使用超链接、引用、参见、相关、其他,并尽量与文本表达保持一致。'
- - '若引用关系带有特定上下文、说明语或引用目的,可写入 context;否则留空。'
- - '不要把仅仅共同出现、主题相近但没有显式互指的节点误判为交叉引用关系。'
-
-identifiers:
- entity_id: name
- relation_id: '{source}|{relation_type}|{target}'
- relation_members:
- source: source
- target: target
-
-options:
- extraction_mode: two_stage
-
-display:
- entity_label: '{name} ({type})'
- relation_label: '{relation_type}'
diff --git a/hyperextract/templates/presets/general/doc_structure.yaml b/hyperextract/templates/presets/general/doc_structure.yaml
new file mode 100644
index 0000000..5fcd66c
--- /dev/null
+++ b/hyperextract/templates/presets/general/doc_structure.yaml
@@ -0,0 +1,110 @@
+language: [zh, en]
+
+name: doc_structure
+type: graph
+tags: [general, document, structure, knowledge]
+
+description:
+ zh: '文档结构图模板 - 从技术文档、论文、报告中提取章节层级结构和交叉引用关系'
+ en: 'Document Structure Template - Extract chapter hierarchy and cross-reference relationships from technical documents, papers, and reports'
+
+output:
+ description:
+ zh: '文档的章节层级结构和引用关系网络'
+ en: 'Document chapter hierarchy and reference relationship network'
+ entities:
+ description:
+ zh: '文档中的章节和内容节点'
+ en: 'Chapter and content nodes in the document'
+ fields:
+ - name: title
+ type: str
+ description:
+ zh: '章节标题或内容标识'
+ en: 'Chapter title or content identifier'
+ required: true
+ - name: node_type
+ type: str
+ description:
+ zh: '节点类型:章/节/小节/段落/代码/表格/图表'
+ en: 'Node type: chapter/section/subsection/paragraph/code/table/figure'
+ required: true
+ - name: level
+ type: int
+ description:
+ zh: '层级深度,从1开始(1=章,2=节,3=小节...)'
+ en: 'Hierarchy level starting from 1 (1=chapter, 2=section, 3=subsection...)'
+ required: true
+ - name: summary
+ type: str
+ description:
+ zh: '章节内容摘要'
+ en: 'Chapter content summary'
+ required: false
+ relations:
+ description:
+ zh: '章节之间的包含关系和交叉引用'
+ en: 'Containment relationships and cross-references between chapters'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '源节点标题'
+ en: 'Source node title'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '目标节点标题'
+ en: 'Target node title'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型,用简洁词语概括两个节点之间的连接方式'
+ en: 'Relation type, concisely describing the connection between two nodes'
+ required: true
+ - name: reference_context
+ type: str
+ description:
+ zh: '引用上下文说明(如"如图3.2所示")'
+ en: 'Reference context (e.g., "as shown in Figure 3.2")'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位技术文档分析师,擅长从技术文档、论文和报告中提取结构化的章节层级和引用关系。'
+ en: 'You are a technical document analyst skilled at extracting structured chapter hierarchy and reference relationships from technical documents, papers, and reports.'
+ rules_for_entities:
+ zh:
+ - '提取所有具有独立标题的章节作为节点'
+ - 'code类型节点应标注代码片段的主要功能'
+ - 'table和figure类型节点应标注表格/图表的主题'
+ - 'level值应与文档实际层级结构一致'
+ - 'summary应概括该章节的核心内容,不超过50字'
+ en:
+ - 'Extract all chapters with independent titles as nodes'
+ - 'code type nodes should indicate the main functionality of code snippets'
+ - 'table and figure type nodes should indicate the subject of the table/figure'
+ - 'level values should match the actual document hierarchy'
+ - 'summary should summarize the core content of the chapter, not exceeding 50 characters'
+ rules_for_relations:
+ zh:
+ - 'type用简洁词语概括关系性质'
+ - '仅提取文档中明确标记的引用关系(如"见第X章"、"如图X所示")'
+ - '不要创建隐含的语义关联,仅提取显式的引用关系'
+ en:
+ - 'Type should concisely describe the nature of the relationship'
+ - 'Only extract explicitly marked reference relationships (e.g., "see Chapter X", "as shown in Figure X")'
+ - 'Do not create implicit semantic associations, only explicit reference relationships'
+
+identifiers:
+ entity_id: title
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{title}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/general/life_event_timeline.yaml b/hyperextract/templates/presets/general/life_event_timeline.yaml
deleted file mode 100644
index ccc5902..0000000
--- a/hyperextract/templates/presets/general/life_event_timeline.yaml
+++ /dev/null
@@ -1,104 +0,0 @@
-language: zh
-name: biographical_timeline
-type: temporal_graph
-tags: [general]
-description:
- zh: '人物生平时间线 - 从传记、自传、回忆录等人物叙事文本中提取带时间的生平事件,并组织为时间线。'
-
-output:
- description:
- zh: '由人物相关实体和生平事件关系组成的时间线结果,用于还原人物经历、角色变化与关键转折。'
- entities:
- description:
- zh: '文本中与人物生平相关、可作为事件端点的实体。'
- fields:
- - name: name
- type: str
- description:
- zh: '实体在文本中的稳定名称或常用称呼。'
- - name: type
- type: str
- description:
- zh: '实体类型,如人物、地点、组织、作品、奖项、身份、物品、概念、其他。'
- - name: description
- type: str
- description:
- zh: '对实体身份、角色或在当前文本中作用的简要说明。'
- required: false
- relations:
- description:
- zh: '围绕人物经历构建的事件关系边。'
- fields:
- - name: source
- type: str
- description:
- zh: '事件发起者或经历者,通常为人物主体或触发事件的相关实体。'
- - name: target
- type: str
- description:
- zh: '与该事件直接相关的对象实体,如地点、学校、组织、人物、作品、奖项等。'
- - name: eventType
- type: str
- description:
- zh: '事件类型。使用能够概括事件本身的简洁词语或短语,优先具体、直观、可区分;通常可用动词,如出生、入学、任职、结婚、出版、获奖。可参考示例,但不必局限于示例,也可以根据文本语义自行创造更贴切的标签。'
- - name: eventDate
- type: str
- description:
- zh: '事件时间。优先保留文本中最精确的时间表达;可为年份、年月、年月日,必要时可保留时间段。'
- required: false
- - name: details
- type: str
- description:
- zh: '事件补充说明,可记录身份、地点、机构、结果或上下文。'
- required: false
-
-guideline:
- target:
- zh: '你是一位专业的人物传记事件抽取专家。请围绕人物主体,从文本中提取与其生平经历直接相关的实体和事件,并构建可信的时间线。'
- rules_for_entities:
- zh:
- - '优先提取与人物主体生平经历直接相关的实体,如人物、地点、组织、作品、奖项、身份称谓等。'
- - '为每个实体指定尽量准确的类型,并优先使用文本中的稳定称呼。'
- - '不必为纯背景概念、空泛评价或无法成为事件端点的内容单独建实体。'
- rules_for_relations:
- zh:
- - '仅从已提取实体中创建事件边,source 和 target 都必须对应实体 name。'
- - '事件应代表可放入人物时间线的经历、转折、公开活动或身份变化,不要把纯评价、性格描述或静态背景知识当作事件。'
- - '若一句话包含多个独立事件,应拆分为多条关系;若同一事件同时涉及时间、地点、组织、身份,可在 details 中补充。'
- - 'eventType 应使用能够概括事件本身的简洁词语或短语,优先具体、直观、可区分;通常可用动词或动宾结构。'
- - '优先写贴近原文语义的具体标签,不要过度抽象。例如优先写“入学、毕业、留学”而不是“教育”,优先写“任职、升迁、离任”而不是“工作”,优先写“出版、获奖、发明”而不是“成就”。'
- - '可参考但不限于以下示例:出生、逝世、迁居、入学、毕业、任职、结婚、合作、出版、获奖、参战、归国、隐居。'
- - '允许根据文本语义自行创造更贴切的标签,例如拜师、出道、登基、封侯、退赛、复出等;不要求必须使用示例中的词。'
- - '仅在确实无法提炼出更具体的事件标签时,才使用较泛的词。'
- - '时间信息为可选,仅在文本明确给出或能根据明确相对时间可靠换算时才填写。'
- rules_for_time:
- zh:
- - '当前观察日期:{observation_time}'
- - '相对时间解析(基于观察日期):'
- - '- "去年" → {observation_time} 前一年'
- - '- "今年" → {observation_time} 所属年份'
- - '- "明年" → {observation_time} 后一年'
- - '- "昨天/今天/明天" → {observation_time} 前/当/后一天'
- - '- "十年前/五年后" → {observation_time} 前/后相应年数'
- - '- "年初/年末" → 当年 1月1日/12月31日'
- - '- "朝代" → 保留朝代名称'
- - '明确时间:优先保留文本中最精确的原始格式'
- - '- 年份(如 429年、2023年)→ 429、2023'
- - '- 年-月、年-月-日、时间段(如 1937-1945)→ 保持原格式'
- - '若只有模糊阶段性表达且无法可靠换算,不要强行补时间'
- - '缺失时间:留空'
-
-identifiers:
- entity_id: name
- relation_id: '{source}|{target}'
- relation_members:
- source: source
- target: target
- time_field: 'eventDate'
-
-options:
- extraction_mode: two_stage
-
-display:
- entity_label: '{name} ({type})'
- relation_label: '{eventType}({eventDate})'
diff --git a/hyperextract/templates/presets/general/social_network.yaml b/hyperextract/templates/presets/general/social_network.yaml
deleted file mode 100644
index cfdc3b5..0000000
--- a/hyperextract/templates/presets/general/social_network.yaml
+++ /dev/null
@@ -1,87 +0,0 @@
-language: zh
-name: social_network
-type: graph
-tags: [general]
-description:
- zh: '社会关系网络模板 - 从人物叙事、组织介绍、历史材料或新闻文本中提取人物与组织之间的人际和隶属关系。'
-
-output:
- description:
- zh: '由人物、组织等节点及其关系边组成的社会关系网络结构。'
- entities:
- description:
- zh: '人际关系网络中的人物、组织或机构节点。'
- fields:
- - name: name
- type: str
- description:
- zh: '人物、组织或机构的名称。'
- - name: type
- type: str
- description:
- zh: '节点类型,如人物、组织、机构、团体、其他。'
- - name: description
- type: str
- description:
- zh: '对节点身份或其在网络中作用的简要说明。'
- required: false
- default: ''
- - name: role
- type: str
- description:
- zh: '该节点在当前文本中的角色、身份或职位。'
- required: false
- default: ''
- relations:
- description:
- zh: '人物与人物、人物与组织、组织与组织之间的社会关系边。'
- fields:
- - name: source
- type: str
- description:
- zh: '关系起点节点。'
- - name: target
- type: str
- description:
- zh: '关系终点节点。'
- - name: relation_type
- type: str
- description:
- zh: '关系类型,如亲属、朋友、同事、师生、上下级、合作、竞争、隶属、其他。'
- - name: details
- type: str
- description:
- zh: '对关系依据、互动方式或上下文的简要补充。'
- required: false
- default: ''
-
-guideline:
- target:
- zh: '你是一位社会关系分析专家。请从文本中提取人物、组织及其明确存在的人际关系或组织隶属关系。'
- rules_for_entities:
- zh:
- - '优先提取对关系网络有意义的人物、组织、机构或团体节点。'
- - 'name 使用文本中最稳定、最清晰的称呼;同一对象保持统一命名。'
- - 'type 概括节点类别,role 用于补充其在当前文本中的身份或职位;没有则留空。'
- - 'description 可说明该节点为何重要或其在关系网络中的作用;信息不足时留空。'
- rules_for_relations:
- zh:
- - '仅在文本明确表达两节点之间存在社会关系、合作关系、竞争关系或组织隶属时建立边。'
- - 'source 和 target 必须直接引用已提取节点的 name。'
- - 'relation_type 优先使用亲属、朋友、同事、师生、上下级、合作、竞争、隶属、其他。'
- - 'details 可用于补充关系背景、时间段或互动情境;没有额外信息时留空。'
- - '不要把共同出现、共同被提及但无明确互动的对象误判为存在关系。'
-
-identifiers:
- entity_id: name
- relation_id: '{source}|{relation_type}|{target}'
- relation_members:
- source: source
- target: target
-
-options:
- extraction_mode: two_stage
-
-display:
- entity_label: '{name} ({role})'
- relation_label: '{relation_type}'
diff --git a/hyperextract/templates/presets/general/workflow_graph.yaml b/hyperextract/templates/presets/general/workflow_graph.yaml
new file mode 100644
index 0000000..2e897eb
--- /dev/null
+++ b/hyperextract/templates/presets/general/workflow_graph.yaml
@@ -0,0 +1,122 @@
+language: [zh, en]
+
+name: workflow_graph
+type: temporal_graph
+tags: [general, workflow, process, agent, skill]
+
+description:
+ zh: '工作流程图模板 - 从Skill定义文件、Agent工作流文档、SOP操作手册中提取流程步骤及其执行顺序、依赖关系和条件分支'
+ en: 'Workflow Graph Template - Extract workflow steps, execution order, dependencies, and conditional branches from Skill definitions, Agent workflow documents, and SOP manuals'
+
+output:
+ description:
+ zh: '工作流程的步骤节点及其执行关系'
+ en: 'Workflow steps and their execution relationships'
+ entities:
+ description:
+ zh: '工作流程中的步骤节点'
+ en: 'Step nodes in the workflow'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '步骤名称,使用清晰的动作性描述'
+ en: 'Step name with clear action-oriented description'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '步骤类型:调用技能/普通步骤/条件判断/分支节点'
+ en: 'Step type: skill/step/condition/branch'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '步骤的简要功能描述'
+ en: 'Brief description of step functionality'
+ required: false
+ - name: input
+ type: str
+ description:
+ zh: '步骤的输入要求'
+ en: 'Input requirements for the step'
+ required: false
+ - name: output
+ type: str
+ description:
+ zh: '步骤的输出结果'
+ en: 'Output result of the step'
+ required: false
+ relations:
+ description:
+ zh: '步骤之间的执行顺序、依赖关系和条件分支'
+ en: 'Execution order, dependencies, and conditional branches between steps'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '源步骤名称'
+ en: 'Source step name'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '目标步骤名称'
+ en: 'Target step name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型,用简洁词语概括两个步骤之间的连接方式'
+ en: 'Relation type, concisely describing the connection between two steps'
+ required: true
+ - name: condition
+ type: str
+ description:
+ zh: '条件表达式(如分支条件)'
+ en: 'Condition expression (e.g., branch condition)'
+ required: false
+ - name: description
+ type: str
+ description:
+ zh: '关系说明,如分支条件或触发原因'
+ en: 'Relation description such as branch condition or trigger reason'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位工作流分析师,擅长从Skill定义文档、Agent工作流说明和SOP操作手册中提取流程结构。'
+ en: 'You are a workflow analyst skilled at extracting process structures from Skill definitions, Agent workflow documentation, and SOP manuals.'
+ rules_for_entities:
+ zh:
+ - '提取每个具有独立功能的步骤作为节点'
+ - '同一步骤在文档中保持命名一致'
+ - '仅提取文档中明确描述的步骤'
+ - 'input/output应简洁描述数据流向'
+ en:
+ - 'Extract each functionally independent step as a node'
+ - 'Maintain consistent naming for the same step throughout the document'
+ - 'Only extract steps explicitly described in the document'
+ - 'input/output should concisely describe data flow'
+ rules_for_relations:
+ zh:
+ - '仅在文档明确表达两个步骤之间存在关系时创建关系'
+ - '优先使用文档中明确出现的关系词'
+ - '不要创建隐含的因果关系'
+ - '并行执行可用相应关系类型表示'
+ en:
+ - 'Create relations only when the document explicitly expresses a relationship between two steps'
+ - 'Prefer relation words that explicitly appear in the document'
+ - 'Do not create implicit causal relationships'
+ - 'Parallel execution can be represented using appropriate relation type'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/industry/.gitkeep b/hyperextract/templates/presets/industry/.gitkeep
deleted file mode 100644
index e69de29..0000000
diff --git a/hyperextract/templates/presets/industry/emergency_response.yaml b/hyperextract/templates/presets/industry/emergency_response.yaml
new file mode 100644
index 0000000..9a3a417
--- /dev/null
+++ b/hyperextract/templates/presets/industry/emergency_response.yaml
@@ -0,0 +1,100 @@
+language: [zh, en]
+
+name: emergency_response
+type: graph
+tags: [industry, emergency, response, safety]
+
+description:
+ zh: '应急响应图模板 - 从应急预案中提取事故场景、响应动作和责任人的关系'
+ en: 'Emergency Response Graph Template - Extract relationships between accident scenarios, response actions, and responsible persons from emergency plans'
+
+output:
+ description:
+ zh: '事故到响应动作到责任人的触发链'
+ en: 'Accident to response action to responsible person trigger chain'
+ entities:
+ description:
+ zh: '应急响应要素'
+ en: 'Emergency response elements'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '要素名称'
+ en: 'Element name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '要素类型:事故场景/响应动作/资源'
+ en: 'Element type: accident scenario/response action/resource'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '说明'
+ en: 'Description'
+ required: false
+ relations:
+ description:
+ zh: '事故、响应动作和责任人的关系'
+ en: 'Relationships between accidents, response actions, and responsible persons'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '源要素'
+ en: 'Source element'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '目标要素'
+ en: 'Target element'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:triggers/executed_by/requires'
+ en: 'Relation type: triggers/executed_by/requires'
+ required: true
+ - name: responsible_role
+ type: str
+ description:
+ zh: '责任角色'
+ en: 'Responsible role'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的应急管理专家,负责从应急预案中准确提取应急响应关系。'
+ en: 'You are a professional emergency management expert responsible for accurately extracting emergency response relationships from emergency plans.'
+ rules_for_entities:
+ zh:
+ - '提取事故场景作为实体'
+ - '提取响应动作作为实体'
+ - '提取所需资源作为实体'
+ en:
+ - 'Extract accident scenarios as entities'
+ - 'Extract response actions as entities'
+ - 'Extract required resources as entities'
+ rules_for_relations:
+ zh:
+ - 'type为triggers表示触发响应'
+ - 'type为executed_by表示执行责任人'
+ - 'type为requires表示所需资源'
+ en:
+ - 'type triggers indicates triggering response'
+ - 'type executed_by indicates executing responsible person'
+ - 'type requires indicates required resources'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/industry/equipment_topology.yaml b/hyperextract/templates/presets/industry/equipment_topology.yaml
new file mode 100644
index 0000000..f618c1c
--- /dev/null
+++ b/hyperextract/templates/presets/industry/equipment_topology.yaml
@@ -0,0 +1,98 @@
+language: [zh, en]
+
+name: equipment_topology
+type: graph
+tags: [industry, equipment, topology, system]
+
+description:
+ zh: '设备拓扑图模板 - 从设备手册中提取设备之间的物理连接和层级关系'
+ en: 'Equipment Topology Graph Template - Extract physical connections and hierarchical relationships between equipment from equipment manuals'
+
+output:
+ description:
+ zh: '设备之间的物理连接和层级关系'
+ en: 'Physical connections and hierarchical relationships between equipment'
+ entities:
+ description:
+ zh: '设备实体'
+ en: 'Equipment entities'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '设备名称'
+ en: 'Equipment name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '设备类型'
+ en: 'Equipment type'
+ required: true
+ - name: location
+ type: str
+ description:
+ zh: '安装位置'
+ en: 'Installation location'
+ required: false
+ - name: specifications
+ type: str
+ description:
+ zh: '规格参数'
+ en: 'Specifications'
+ required: false
+ relations:
+ description:
+ zh: '设备之间的连接和层级关系'
+ en: 'Connection and hierarchical relationships between equipment'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '源设备'
+ en: 'Source equipment'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '目标设备'
+ en: 'Target equipment'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:connected_to/located_in/contains'
+ en: 'Relation type: connected_to/located_in/contains'
+ required: true
+
+guideline:
+ target:
+ zh: '你是一位专业的设备工程师,负责从设备手册中准确提取设备拓扑关系。'
+ en: 'You are a professional equipment engineer responsible for accurately extracting equipment topology relationships from equipment manuals.'
+ rules_for_entities:
+ zh:
+ - '提取设备名称作为实体'
+ - '设备类型应准确分类'
+ - '安装位置应标注具体的物理位置'
+ en:
+ - 'Extract equipment names as entities'
+ - 'Equipment types should be accurately classified'
+ - 'Installation locations should indicate specific physical locations'
+ rules_for_relations:
+ zh:
+ - '优先提取设备手册中明确描述的设备关系'
+ - '仅提取文本中明确表达的设备关系'
+ en:
+ - 'Prioritize explicitly described equipment relationships from equipment manuals'
+ - 'Only extract equipment relationships explicitly expressed in the text'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/industry/failure_case.yaml b/hyperextract/templates/presets/industry/failure_case.yaml
new file mode 100644
index 0000000..965919d
--- /dev/null
+++ b/hyperextract/templates/presets/industry/failure_case.yaml
@@ -0,0 +1,102 @@
+language: [zh, en]
+
+name: failure_case
+type: graph
+tags: [industry, failure, maintenance, troubleshooting]
+
+description:
+ zh: '故障案例图模板 - 从故障分析报告中提取故障现象、原因和措施的因果关系'
+ en: 'Failure Case Graph Template - Extract failure phenomenon-cause-measure causal relationships from failure analysis reports'
+
+output:
+ description:
+ zh: '故障现象到原因到措施的因果链'
+ en: 'Failure phenomenon to cause to measure causal chain'
+ entities:
+ description:
+ zh: '故障要素'
+ en: 'Failure elements'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '要素名称'
+ en: 'Element name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '要素类型:故障现象/原因/措施'
+ en: 'Element type: phenomenon/cause/measure'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '说明'
+ en: 'Description'
+ required: false
+ relations:
+ description:
+ zh: '故障要素之间的因果关系'
+ en: 'Causal relationships between failure elements'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '前置要素'
+ en: 'Prior element'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '后续要素'
+ en: 'Subsequent element'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:caused_by/resolved_by/prevented_by'
+ en: 'Relation type: caused_by/resolved_by/prevented_by'
+ required: true
+ - name: lesson
+ type: str
+ description:
+ zh: '经验教训'
+ en: 'Lessons learned'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的设备工程师,负责从故障分析报告中准确提取故障因果关系。'
+ en: 'You are a professional equipment engineer responsible for accurately extracting failure causal relationships from failure analysis reports.'
+ rules_for_entities:
+ zh:
+ - '提取故障现象作为实体'
+ - '提取根本原因作为实体'
+ - '提取纠正措施作为实体'
+ en:
+ - 'Extract failure phenomena as entities'
+ - 'Extract root causes as entities'
+ - 'Extract corrective measures as entities'
+ rules_for_relations:
+ zh:
+ - 'type为caused_by表示由某原因引起'
+ - 'type为resolved_by表示通过某措施解决'
+ - 'type为prevented_by表示通过某措施预防'
+ - 'lesson字段应总结经验教训'
+ en:
+ - 'type caused_by indicates caused by a reason'
+ - 'type resolved_by indicates resolved by a measure'
+ - 'type prevented_by indicates prevented by a measure'
+ - 'Lesson field should summarize lessons learned'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/industry/operation_flow.yaml b/hyperextract/templates/presets/industry/operation_flow.yaml
new file mode 100644
index 0000000..5170e40
--- /dev/null
+++ b/hyperextract/templates/presets/industry/operation_flow.yaml
@@ -0,0 +1,108 @@
+language: [zh, en]
+
+name: operation_flow
+type: graph
+tags: [industry, operation, sop, workflow]
+
+description:
+ zh: '操作流程图模板 - 从SOP操作手册中提取操作步骤和流程关系'
+ en: 'Operation Flow Graph Template - Extract operation steps and flow relationships from SOP manuals'
+
+output:
+ description:
+ zh: '操作步骤的流程关系'
+ en: 'Operation step flow relationships'
+ entities:
+ description:
+ zh: '操作步骤'
+ en: 'Operation steps'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '步骤名称'
+ en: 'Step name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '步骤类型:动作/检查/切换'
+ en: 'Step type: action/inspection/switch'
+ required: true
+ - name: equipment_state
+ type: str
+ description:
+ zh: '设备状态'
+ en: 'Equipment state'
+ required: false
+ - name: expected_result
+ type: str
+ description:
+ zh: '预期结果'
+ en: 'Expected result'
+ required: false
+ relations:
+ description:
+ zh: '步骤之间的执行关系'
+ en: 'Execution relationships between steps'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '前置步骤'
+ en: 'Prior step'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '后续步骤'
+ en: 'Subsequent step'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:next/triggers/results_in'
+ en: 'Relation type: next/triggers/results_in'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '关系说明'
+ en: 'Relationship description'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的操作工程师,负责从SOP操作手册中准确提取操作流程关系。'
+ en: 'You are a professional operation engineer responsible for accurately extracting operation flow relationships from SOP manuals.'
+ rules_for_entities:
+ zh:
+ - '提取每个具有独立功能的操作步骤作为实体'
+ - '步骤类型应准确分类'
+ - '设备状态应标注关键状态变化'
+ en:
+ - 'Extract each operation step with independent function as entity'
+ - 'Step types should be accurately classified'
+ - 'Equipment states should indicate key state changes'
+ rules_for_relations:
+ zh:
+ - 'type为next表示顺序执行'
+ - 'type为triggers表示触发条件'
+ - 'type为results_in表示导致状态变化'
+ - '仅提取文档中明确说明的关系'
+ en:
+ - 'type next indicates sequential execution'
+ - 'type triggers indicates trigger condition'
+ - 'type results_in indicates state change'
+ - 'Only extract explicitly stated relationships from the text'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/industry/safety_control.yaml b/hyperextract/templates/presets/industry/safety_control.yaml
new file mode 100644
index 0000000..d29352a
--- /dev/null
+++ b/hyperextract/templates/presets/industry/safety_control.yaml
@@ -0,0 +1,102 @@
+language: [zh, en]
+
+name: safety_control
+type: graph
+tags: [industry, safety, hazard, control]
+
+description:
+ zh: '安全控制图模板 - 从安全管理手册中提取危险源、风险和控制措施的关系'
+ en: 'Safety Control Graph Template - Extract hazard-risk-control relationships from safety management manuals'
+
+output:
+ description:
+ zh: '危险源到风险到控制措施的关系链条'
+ en: 'Hazard to risk to control measure relationship chain'
+ entities:
+ description:
+ zh: '安全要素'
+ en: 'Safety elements'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '要素名称'
+ en: 'Element name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '要素类型:危险源/风险点/控制措施'
+ en: 'Element type: hazard/risk point/control measure'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '说明'
+ en: 'Description'
+ required: false
+ relations:
+ description:
+ zh: '危险源、风险和控制措施的关系'
+ en: 'Relationships between hazards, risks, and control measures'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '源要素'
+ en: 'Source element'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '目标要素'
+ en: 'Target element'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:causes/mitigates/assigned_to'
+ en: 'Relation type: causes/mitigates/assigned_to'
+ required: true
+ - name: responsible_role
+ type: str
+ description:
+ zh: '责任角色'
+ en: 'Responsible role'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的安全工程师,负责从安全管理手册中准确提取安全控制关系。'
+ en: 'You are a professional safety engineer responsible for accurately extracting safety control relationships from safety management manuals.'
+ rules_for_entities:
+ zh:
+ - '提取危险源作为实体'
+ - '提取风险点作为实体'
+ - '提取控制措施作为实体'
+ en:
+ - 'Extract hazards as entities'
+ - 'Extract risk points as entities'
+ - 'Extract control measures as entities'
+ rules_for_relations:
+ zh:
+ - 'type为causes表示导致风险'
+ - 'type为mitigates表示减轻风险'
+ - 'type为assigned_to表示责任分配'
+ - 'responsible_role应标注具体的责任人或部门'
+ en:
+ - 'type causes indicates causing risk'
+ - 'type mitigates indicates mitigating risk'
+ - 'type assigned_to indicates responsibility assignment'
+ - 'Responsible role should indicate specific responsible person or department'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/legal/.gitkeep b/hyperextract/templates/presets/legal/.gitkeep
deleted file mode 100644
index e69de29..0000000
diff --git a/hyperextract/templates/presets/legal/case_citation.yaml b/hyperextract/templates/presets/legal/case_citation.yaml
new file mode 100644
index 0000000..94f45bc
--- /dev/null
+++ b/hyperextract/templates/presets/legal/case_citation.yaml
@@ -0,0 +1,102 @@
+language: [zh, en]
+
+name: case_citation
+type: graph
+tags: [legal, case, citation, precedent]
+
+description:
+ zh: '判例引用图模板 - 从判决书和法律意见书中提取判例引用关系'
+ en: 'Case Citation Graph Template - Extract case citation relationships from judgments and legal opinions'
+
+output:
+ description:
+ zh: '判例之间的引用关系网络'
+ en: 'Citation relationship network between cases'
+ entities:
+ description:
+ zh: '案件和引用条款'
+ en: 'Cases and cited provisions'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '案件名称或条款编号'
+ en: 'Case name or provision number'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '实体类型:案件/条款/法规'
+ en: 'Entity type: case/provision/regulation'
+ required: true
+ - name: court
+ type: str
+ description:
+ zh: '审理法院'
+ en: 'Adjudicating court'
+ required: false
+ relations:
+ description:
+ zh: '案件之间的引用和效力关系'
+ en: 'Citation and authority relationships between cases'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '引用方'
+ en: 'Citing case'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '被引用方'
+ en: 'Cited case'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:cites/distinguished_from/overruled_by'
+ en: 'Relation type: cites/distinguished_from/overruled_by'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '引用说明'
+ en: 'Citation description'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的法律研究员,负责从判决书和法律意见书中准确提取判例引用关系。'
+ en: 'You are a professional legal researcher responsible for accurately extracting case citation relationships from judgments and legal opinions.'
+ rules_for_entities:
+ zh:
+ - '提取案件名称作为核心实体'
+ - '条款实体应包含完整的条款编号'
+ - '法院信息应准确标注'
+ en:
+ - 'Extract case names as core entities'
+ - 'Provision entities should include complete provision numbers'
+ - 'Court information should be accurately labeled'
+ rules_for_relations:
+ zh:
+ - 'type为cites表示引用关系'
+ - 'type为distinguished_from表示区分解释'
+ - 'type为overruled_by表示被推翻'
+ - '仅提取文本中明确提及的引用关系'
+ en:
+ - 'type cites indicates citation relationship'
+ - 'type distinguished_from indicates distinguishing explanation'
+ - 'type overruled_by indicates being overruled'
+ - 'Only extract citation relationships explicitly mentioned in the text'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/legal/case_fact_timeline.yaml b/hyperextract/templates/presets/legal/case_fact_timeline.yaml
new file mode 100644
index 0000000..f495568
--- /dev/null
+++ b/hyperextract/templates/presets/legal/case_fact_timeline.yaml
@@ -0,0 +1,109 @@
+language: [zh, en]
+
+name: case_fact_timeline
+type: temporal_graph
+tags: [legal, case, fact, timeline]
+
+description:
+ zh: '案件事实时间线模板 - 从判决书和证据材料中提取案件事实的时间线'
+ en: 'Case Fact Timeline Template - Extract case fact timelines from judgments and evidence materials'
+
+output:
+ description:
+ zh: '案件事实的时间顺序和证据链'
+ en: 'Temporal sequence of case facts and evidence chain'
+ entities:
+ description:
+ zh: '案件事实和证据'
+ en: 'Case facts and evidence'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '事实描述或证据编号'
+ en: 'Fact description or evidence number'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '实体类型:事实/证据'
+ en: 'Entity type: fact/evidence'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '详细描述'
+ en: 'Detailed description'
+ required: false
+ relations:
+ description:
+ zh: '事实之间的时间顺序和证据支持关系'
+ en: 'Temporal sequence and evidence support relationships between facts'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '前置事实'
+ en: 'Prior fact'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '后续事实'
+ en: 'Subsequent fact'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:preceded_by/proves/disputes'
+ en: 'Relation type: preceded_by/proves/disputes'
+ required: true
+ - name: time
+ type: str
+ description:
+ zh: '事实发生日期'
+ en: 'Fact occurrence date'
+ required: false
+ - name: disputed
+ type: str
+ description:
+ zh: '是否争议:是/否'
+ en: 'Whether disputed: yes/no'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的诉讼分析师,负责从判决书和证据材料中准确提取案件事实并构建时间线。'
+ en: 'You are a professional litigation analyst responsible for accurately extracting case facts and constructing timelines from judgments and evidence materials.'
+ rules_for_entities:
+ zh:
+ - '提取关键的案件事实作为实体'
+ - '证据实体应包含证据编号和来源'
+ - '实体描述应精炼准确'
+ en:
+ - 'Extract key case facts as entities'
+ - 'Evidence entities should include evidence number and source'
+ - 'Entity descriptions should be concise and accurate'
+ rules_for_relations:
+ zh:
+ - 'type为preceded_by表示时间先后顺序'
+ - 'type为proves表示证据支持'
+ - 'type为disputes表示对某一事实存在争议'
+ - '仅提取文本中明确陈述的事实关系'
+ en:
+ - 'type preceded_by indicates temporal sequence'
+ - 'type proves indicates evidence support'
+ - 'type disputes indicates dispute over a fact'
+ - 'Only extract explicitly stated fact relationships from the text'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+ time_field: time
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/legal/compliance_list.yaml b/hyperextract/templates/presets/legal/compliance_list.yaml
new file mode 100644
index 0000000..2ac2649
--- /dev/null
+++ b/hyperextract/templates/presets/legal/compliance_list.yaml
@@ -0,0 +1,60 @@
+language: [zh, en]
+
+name: compliance_list
+type: list
+tags: [legal, compliance, regulation]
+
+description:
+ zh: '合规要求清单模板 - 从合规手册和审计报告中提取结构化的合规要求'
+ en: 'Compliance Requirements List Template - Extract structured compliance requirements from compliance manuals and audit reports'
+
+output:
+ description:
+ zh: '合规要求和监管事项清单'
+ en: 'List of compliance requirements and regulatory items'
+ fields:
+ - name: requirement
+ type: str
+ description:
+ zh: '合规要求描述'
+ en: 'Compliance requirement description'
+ required: true
+ - name: regulation
+ type: str
+ description:
+ zh: '依据法规'
+ en: 'Basis regulation'
+ required: true
+ - name: deadline
+ type: str
+ description:
+ zh: '完成期限'
+ en: 'Completion deadline'
+ required: false
+ - name: responsible_party
+ type: str
+ description:
+ zh: '责任人或责任部门'
+ en: 'Responsible party or department'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的合规分析师,负责从合规文档中准确提取结构化的合规要求。'
+ en: 'You are a professional compliance analyst responsible for accurately extracting structured compliance requirements from compliance documents.'
+ rules:
+ zh:
+ - '提取的合规要求应完整保留原文的核心含义'
+ - '依据法规应标注具体的法规名称和条款'
+ - '完成期限应明确标注时间节点'
+ - '责任人应尽可能具体到部门或个人'
+ en:
+ - 'Extracted compliance requirements should fully preserve the core meaning from the source text'
+ - 'Basis regulations should specify the specific regulation name and clause'
+ - 'Completion deadlines should clearly indicate time milestones'
+ - 'Responsible parties should be as specific as possible to department or individual'
+
+identifiers: {}
+
+display:
+ label: '{requirement}'
diff --git a/hyperextract/templates/presets/legal/contract_obligation.yaml b/hyperextract/templates/presets/legal/contract_obligation.yaml
new file mode 100644
index 0000000..d7ef394
--- /dev/null
+++ b/hyperextract/templates/presets/legal/contract_obligation.yaml
@@ -0,0 +1,100 @@
+language: [zh, en]
+
+name: contract_obligation
+type: hypergraph
+tags: [legal, contract, obligation, parties]
+
+description:
+ zh: '合同义务图模板 - 从服务合同和采购合同中提取各方的权利义务关系'
+ en: 'Contract Obligation Graph Template - Extract rights and obligations of all parties from service contracts and procurement contracts'
+
+output:
+ description:
+ zh: '合同中各方的权利义务关系网络'
+ en: 'Network of rights and obligations of all parties in the contract'
+ entities:
+ description:
+ zh: '合同主体和义务要素'
+ en: 'Contract parties and obligation elements'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '主体名称或义务要素'
+ en: 'Party name or obligation element'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '实体类型:当事人/义务/条件/例外'
+ en: 'Entity type: party/obligation/condition/exception'
+ required: true
+ relations:
+ description:
+ zh: '合同义务的触发和履行关系'
+ en: 'Triggering and performance relationships of contract obligations'
+ fields:
+ - name: party
+ type: list
+ description:
+ zh: '涉及的一方或多方当事人'
+ en: 'One or more parties involved'
+ required: true
+ - name: obligation
+ type: str
+ description:
+ zh: '具体义务内容'
+ en: 'Specific obligation content'
+ required: true
+ - name: trigger
+ type: str
+ description:
+ zh: '义务触发条件'
+ en: 'Obligation trigger condition'
+ required: false
+ - name: exception
+ type: str
+ description:
+ zh: '例外情形'
+ en: 'Exception situations'
+ required: false
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:must_perform/triggered_by/exempt_by'
+ en: 'Relation type: must_perform/triggered_by/exempt_by'
+ required: true
+
+guideline:
+ target:
+ zh: '你是一位专业的合同律师,负责从服务合同和采购合同中准确提取各方的权利义务关系。'
+ en: 'You are a professional contract lawyer responsible for accurately extracting rights and obligations of all parties from service contracts and procurement contracts.'
+ rules_for_entities:
+ zh:
+ - '提取所有合同当事人作为实体'
+ - '提取核心义务要素作为实体'
+ - '条件要素和例外情形应单独标注'
+ en:
+ - 'Extract all contract parties as entities'
+ - 'Extract core obligation elements as entities'
+ - 'Condition elements and exceptions should be separately labeled'
+ rules_for_relations:
+ zh:
+ - '优先提取文本中明确描述的义务关系'
+ - '义务触发条件和例外情形应单独标注'
+ - 'party字段应包含所有涉及的当事人'
+ - '仅提取文本中明确表达的义务关系'
+ en:
+ - 'Prioritize explicitly described obligation relationships in the text'
+ - 'Obligation trigger conditions and exception situations should be separately labeled'
+ - 'Party field should include all parties involved'
+ - 'Only extract obligation relationships explicitly expressed in the text'
+
+identifiers:
+ entity_id: name
+ relation_id: '{obligation}|{type}'
+ relation_members: party
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/legal/defined_term_set.yaml b/hyperextract/templates/presets/legal/defined_term_set.yaml
new file mode 100644
index 0000000..9204afd
--- /dev/null
+++ b/hyperextract/templates/presets/legal/defined_term_set.yaml
@@ -0,0 +1,55 @@
+language: [zh, en]
+
+name: defined_term_set
+type: set
+tags: [legal, contract, terminology]
+
+description:
+ zh: '定义术语集合模板 - 从合同和法律文书中提取术语定义'
+ en: 'Defined Term Set Template - Extract term definitions from contracts and legal documents'
+
+output:
+ description:
+ zh: '合同中的术语定义集合'
+ en: 'Set of term definitions in contracts'
+ fields:
+ - name: term
+ type: str
+ description:
+ zh: '被定义的术语'
+ en: 'Defined term'
+ required: true
+ - name: defined_in
+ type: str
+ description:
+ zh: '定义所在位置(如条款编号)'
+ en: 'Location of definition (e.g., clause number)'
+ required: true
+ - name: meaning
+ type: str
+ description:
+ zh: '术语含义摘要'
+ en: 'Summary of term meaning'
+ required: true
+
+guideline:
+ target:
+ zh: '你是一位专业的法律文书分析师,负责从合同中准确提取术语定义。'
+ en: 'You are a professional legal document analyst responsible for accurately extracting term definitions from contracts.'
+ rules:
+ zh:
+ - '提取所有明确定义的术语'
+ - '定义位置应标注具体的条款或章节'
+ - '含义摘要应精炼概括核心定义'
+ - '同一术语的多次定义应合并'
+ en:
+ - 'Extract all explicitly defined terms'
+ - 'Definition location should specify the specific clause or section'
+ - 'Meaning summary should concisely summarize the core definition'
+ - 'Multiple definitions of the same term should be merged'
+
+identifiers:
+ item_id: term
+
+display:
+ label: '{term}'
diff --git a/hyperextract/templates/presets/medicine/.gitkeep b/hyperextract/templates/presets/medicine/.gitkeep
deleted file mode 100644
index e69de29..0000000
diff --git a/hyperextract/templates/presets/medicine/anatomy_graph.yaml b/hyperextract/templates/presets/medicine/anatomy_graph.yaml
new file mode 100644
index 0000000..5c598f6
--- /dev/null
+++ b/hyperextract/templates/presets/medicine/anatomy_graph.yaml
@@ -0,0 +1,94 @@
+language: [zh, en]
+
+name: anatomy_graph
+type: graph
+tags: [medicine, anatomy, structure, body]
+
+description:
+ zh: '解剖结构图模板 - 从解剖学教材中提取人体结构的层级关系'
+ en: 'Anatomy Structure Graph Template - Extract hierarchical relationships of human body structures from anatomy textbooks'
+
+output:
+ description:
+ zh: '人体结构的层级包含关系'
+ en: 'Hierarchical containment relationships of human body structures'
+ entities:
+ description:
+ zh: '解剖结构实体'
+ en: 'Anatomical structure entities'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '解剖结构名称'
+ en: 'Anatomical structure name'
+ required: true
+ - name: level
+ type: str
+ description:
+ zh: '结构层级:系统/器官/组织/细胞'
+ en: 'Structure level: system/organ/tissue/cell'
+ required: true
+ - name: location
+ type: str
+ description:
+ zh: '空间位置'
+ en: 'Spatial location'
+ required: false
+ relations:
+ description:
+ zh: '解剖结构之间的包含和位置关系'
+ en: 'Containment and location relationships between anatomical structures'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '父级结构'
+ en: 'Parent structure'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '子级结构'
+ en: 'Child structure'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:contains/located_in'
+ en: 'Relation type: contains/located_in'
+ required: true
+
+guideline:
+ target:
+ zh: '你是一位专业的解剖学专家,负责从解剖学教材中准确提取人体结构的层级关系。'
+ en: 'You are a professional anatomy expert responsible for accurately extracting hierarchical relationships of human body structures from anatomy textbooks.'
+ rules_for_entities:
+ zh:
+ - '提取具有明确解剖学定义的实体'
+ - '结构层级应与解剖学标准一致'
+ - '位置信息应准确描述'
+ en:
+ - 'Extract entities with clear anatomical definitions'
+ - 'Structure levels should be consistent with anatomical standards'
+ - 'Location information should be accurately described'
+ rules_for_relations:
+ zh:
+ - '优先提取解剖学教材中明确描述的结构关系'
+ - '层级关系应遵循解剖学分类体系'
+ - '仅提取文本中明确表达的结构关系'
+ en:
+ - 'Prioritize explicitly described structural relationships from anatomy textbooks'
+ - 'Hierarchical relationships should follow anatomical classification system'
+ - 'Only extract structural relationships explicitly expressed in the text'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/medicine/discharge_instruction.yaml b/hyperextract/templates/presets/medicine/discharge_instruction.yaml
new file mode 100644
index 0000000..6a1c65f
--- /dev/null
+++ b/hyperextract/templates/presets/medicine/discharge_instruction.yaml
@@ -0,0 +1,60 @@
+language: [zh, en]
+
+name: discharge_instruction
+type: model
+tags: [medicine, hospital, discharge, patient-care]
+
+description:
+ zh: '出院指导模板 - 从出院小结中提取患者出院后需知的医疗指导信息'
+ en: 'Discharge Instruction Template - Extract post-discharge medical guidance information from discharge summaries'
+
+output:
+ description:
+ zh: '患者出院后医疗指导'
+ en: 'Post-discharge medical guidance for patients'
+ fields:
+ - name: medications
+ type: list
+ description:
+ zh: '用药列表及用法'
+ en: 'Medication list and usage instructions'
+ required: true
+ - name: follow_up
+ type: str
+ description:
+ zh: '随访安排'
+ en: 'Follow-up arrangement'
+ required: true
+ - name: restrictions
+ type: list
+ description:
+ zh: '注意事项'
+ en: 'Things to note'
+ required: false
+ - name: warning_signs
+ type: list
+ description:
+ zh: '警告信号(需立即就医的情况)'
+ en: 'Warning signs (situations requiring immediate medical attention)'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的医疗文档分析师,负责从出院小结中准确提取患者出院后需知的医疗指导信息。'
+ en: 'You are a professional medical document analyst responsible for accurately extracting post-discharge medical guidance information from discharge summaries.'
+ rules:
+ zh:
+ - '用药列表应包含药物名称、剂量和用法'
+ - '随访安排应明确标注时间节点'
+ - '注意事项应包括饮食、运动等生活指导'
+ - '警告信号应列举需立即就医的紧急情况'
+ en:
+ - 'Medication list should include drug name, dosage, and usage'
+ - 'Follow-up arrangement should clearly indicate time milestones'
+ - 'Things to note should include lifestyle guidance such as diet and exercise'
+ - 'Warning signs should list emergency situations requiring immediate medical attention'
+
+identifiers: {}
+
+display:
+ label: '出院指导'
diff --git a/hyperextract/templates/presets/medicine/drug_interaction.yaml b/hyperextract/templates/presets/medicine/drug_interaction.yaml
new file mode 100644
index 0000000..19ec754
--- /dev/null
+++ b/hyperextract/templates/presets/medicine/drug_interaction.yaml
@@ -0,0 +1,108 @@
+language: [zh, en]
+
+name: drug_interaction
+type: graph
+tags: [medicine, drug, interaction, pharmacy]
+
+description:
+ zh: '药物相互作用图模板 - 从药品说明书中提取药物之间的相互作用关系'
+ en: 'Drug Interaction Graph Template - Extract drug-drug interaction relationships from drug labels and interaction manuals'
+
+output:
+ description:
+ zh: '药物相互作用网络'
+ en: 'Drug interaction network'
+ entities:
+ description:
+ zh: '药物实体'
+ en: 'Drug entities'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '药物名称'
+ en: 'Drug name'
+ required: true
+ - name: drug_class
+ type: str
+ description:
+ zh: '药物类别'
+ en: 'Drug class'
+ required: false
+ - name: description
+ type: str
+ description:
+ zh: '药物说明'
+ en: 'Drug description'
+ required: false
+ relations:
+ description:
+ zh: '药物之间的相互作用'
+ en: 'Interactions between drugs'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '药物A名称'
+ en: 'Drug A name'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '药物B名称'
+ en: 'Drug B name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '相互作用类型:协同/拮抗/禁用'
+ en: 'Interaction type: synergistic/antagonistic/contraindicated'
+ required: true
+ - name: mechanism
+ type: str
+ description:
+ zh: '作用机制'
+ en: 'Mechanism of action'
+ required: false
+ - name: severity
+ type: str
+ description:
+ zh: '严重程度:高/中/低'
+ en: 'Severity: high/medium/low'
+ required: true
+
+guideline:
+ target:
+ zh: '你是一位专业的临床药师,负责从药品说明书中准确提取药物相互作用信息。'
+ en: 'You are a professional clinical pharmacist responsible for accurately extracting drug interaction information from drug labels.'
+ rules_for_entities:
+ zh:
+ - '提取药物通用名称作为实体'
+ - '药物类别应准确标注'
+ - '避免重复的药物实体'
+ en:
+ - 'Extract generic drug names as entities'
+ - 'Drug classes should be accurately labeled'
+ - 'Avoid duplicate drug entities'
+ rules_for_relations:
+ zh:
+ - '优先提取药品说明书中明确描述的药物相互作用'
+ - '严重程度应基于临床风险评估'
+ - '避免重复的相互作用记录'
+ - '仅提取文本中明确表达的相互作用'
+ en:
+ - 'Prioritize explicitly described drug interactions from drug labels'
+ - 'Severity should be based on clinical risk assessment'
+ - 'Avoid duplicate interaction records'
+ - 'Only extract interactions explicitly expressed in the text'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/medicine/hospital_timeline.yaml b/hyperextract/templates/presets/medicine/hospital_timeline.yaml
new file mode 100644
index 0000000..b4310b4
--- /dev/null
+++ b/hyperextract/templates/presets/medicine/hospital_timeline.yaml
@@ -0,0 +1,107 @@
+language: [zh, en]
+
+name: hospital_timeline
+type: temporal_graph
+tags: [medicine, hospital, timeline, clinical]
+
+description:
+ zh: '住院病程时间线模板 - 从病程记录和手术记录中提取患者住院期间的事件时间线'
+ en: 'Hospital Timeline Template - Extract patient in-hospital event timelines from clinical notes and surgical records'
+
+output:
+ description:
+ zh: '患者住院期间的事件顺序'
+ en: 'Sequence of patient in-hospital events'
+ entities:
+ description:
+ zh: '医疗事件'
+ en: 'Medical events'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '事件名称'
+ en: 'Event name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '事件类型:入院/检查/手术/用药/会诊/出院'
+ en: 'Event type: admission/examination/surgery/medication/consultation/discharge'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '事件描述'
+ en: 'Event description'
+ required: false
+ relations:
+ description:
+ zh: '事件之间的顺序和结果关系'
+ en: 'Sequence and result relationships between events'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '前置事件'
+ en: 'Prior event'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '后续事件'
+ en: 'Subsequent event'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:follows/results_in'
+ en: 'Relation type: follows/results_in'
+ required: true
+ - name: time
+ type: str
+ description:
+ zh: '事件发生日期'
+ en: 'Event date'
+ required: true
+ - name: performer
+ type: str
+ description:
+ zh: '执行人'
+ en: 'Performer'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的医疗分析师,负责从病程记录中准确提取患者住院期间的事件时间线。'
+ en: 'You are a professional medical analyst responsible for accurately extracting patient in-hospital event timelines from clinical notes.'
+ rules_for_entities:
+ zh:
+ - '提取关键的医疗事件作为实体'
+ - '事件类型应准确分类'
+ - '描述应精炼概括事件内容'
+ en:
+ - 'Extract key medical events as entities'
+ - 'Event types should be accurately classified'
+ - 'Descriptions should concisely summarize event content'
+ rules_for_relations:
+ zh:
+ - 'type为follows表示事件的时序关系'
+ - 'type为results_in表示因果结果关系'
+ - 'performer应标注执行人或科室'
+ en:
+ - 'type follows indicates temporal sequence of events'
+ - 'type results_in indicates causal result relationship'
+ - 'Performer should indicate the executor or department'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}|{time}'
+ relation_members:
+ source: source
+ target: target
+ time_field: time
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/medicine/treatment_map.yaml b/hyperextract/templates/presets/medicine/treatment_map.yaml
new file mode 100644
index 0000000..38d70c3
--- /dev/null
+++ b/hyperextract/templates/presets/medicine/treatment_map.yaml
@@ -0,0 +1,106 @@
+language: [zh, en]
+
+name: treatment_map
+type: hypergraph
+tags: [medicine, treatment, clinical, decision]
+
+description:
+ zh: '治疗方案图模板 - 从临床指南中提取诊断、治疗和结果的映射关系'
+ en: 'Treatment Map Template - Extract diagnosis-treatment-outcome mapping relationships from clinical guidelines'
+
+output:
+ description:
+ zh: '诊断治疗决策路径'
+ en: 'Diagnosis-treatment decision paths'
+ entities:
+ description:
+ zh: '患者特征和治疗要素'
+ en: 'Patient characteristics and treatment elements'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '要素名称'
+ en: 'Element name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '要素类型:患者特征/诊断/治疗/结果'
+ en: 'Element type: patient characteristic/diagnosis/treatment/outcome'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '要素描述'
+ en: 'Element description'
+ required: false
+ relations:
+ description:
+ zh: '诊断、治疗和结果之间的映射关系'
+ en: 'Mapping relationships between diagnosis, treatment, and outcomes'
+ fields:
+ - name: conditions
+ type: list
+ description:
+ zh: '适用条件/患者特征'
+ en: 'Applicable conditions/patient characteristics'
+ required: false
+ - name: diagnosis
+ type: str
+ description:
+ zh: '诊断结果'
+ en: 'Diagnosis result'
+ required: true
+ - name: treatment
+ type: str
+ description:
+ zh: '治疗方案'
+ en: 'Treatment plan'
+ required: true
+ - name: outcome
+ type: str
+ description:
+ zh: '治疗结果'
+ en: 'Treatment outcome'
+ required: false
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:indicates/contraindicated_with'
+ en: 'Relation type: indicates/contraindicated_with'
+ required: true
+
+guideline:
+ target:
+ zh: '你是一位专业的临床专家,负责从临床指南中准确提取治疗方案映射关系。'
+ en: 'You are a professional clinical expert responsible for accurately extracting treatment mapping relationships from clinical guidelines.'
+ rules_for_entities:
+ zh:
+ - '提取患者特征作为实体'
+ - '提取诊断结果作为实体'
+ - '提取治疗方案作为实体'
+ - '提取治疗结果作为实体'
+ en:
+ - 'Extract patient characteristics as entities'
+ - 'Extract diagnosis results as entities'
+ - 'Extract treatment plans as entities'
+ - 'Extract treatment outcomes as entities'
+ rules_for_relations:
+ zh:
+ - 'type为indicates表示适应症'
+ - 'type为contraindicated_with表示禁忌组合'
+ - 'conditions字段应包含所有相关的患者特征/适用条件'
+ en:
+ - 'type indicates indicates indication'
+ - 'type contraindicated_with indicates contraindicated combination'
+ - 'Conditions field should include all relevant patient characteristics/applicable conditions'
+
+identifiers:
+ entity_id: name
+ relation_id: '{diagnosis}|{type}|{treatment}'
+ relation_members: conditions
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/tcm/.gitkeep b/hyperextract/templates/presets/tcm/.gitkeep
deleted file mode 100644
index e69de29..0000000
diff --git a/hyperextract/templates/presets/tcm/formula_composition.yaml b/hyperextract/templates/presets/tcm/formula_composition.yaml
new file mode 100644
index 0000000..5fc3a4d
--- /dev/null
+++ b/hyperextract/templates/presets/tcm/formula_composition.yaml
@@ -0,0 +1,104 @@
+language: [zh, en]
+
+name: formula_composition
+type: hypergraph
+tags: [tcm, formula, prescription, composition]
+
+description:
+ zh: '方剂组成图模板 - 从经典方剂著作中提取方剂的组成和配伍关系'
+ en: 'Formula Composition Graph Template - Extract formula composition and compatibility from classical formula texts'
+
+output:
+ description:
+ zh: '方剂的组成和配伍关系'
+ en: 'Formula composition and compatibility relationships'
+ entities:
+ description:
+ zh: '方剂和药材'
+ en: 'Formulas and herbs'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '方剂或药材名称'
+ en: 'Formula or herb name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '实体类型:方剂/药材'
+ en: 'Entity type: formula/herb'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '说明'
+ en: 'Description'
+ required: false
+ relations:
+ description:
+ zh: '方剂中药材的角色和配伍关系'
+ en: 'Herb roles and compatibility in formulas'
+ fields:
+ - name: formula
+ type: str
+ description:
+ zh: '方剂名称'
+ en: 'Formula name'
+ required: true
+ - name: herb
+ type: str
+ description:
+ zh: '药材名称'
+ en: 'Herb name'
+ required: true
+ - name: role
+ type: str
+ description:
+ zh: '药材角色:君/臣/佐/使'
+ en: 'Herb role: sovereign/minister/assistant/envoy'
+ required: true
+ - name: dosage
+ type: str
+ description:
+ zh: '剂量'
+ en: 'Dosage'
+ required: false
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:contains/antagonizes'
+ en: 'Relation type: contains/antagonizes'
+ required: true
+
+guideline:
+ target:
+ zh: '你是一位专业的中医药学家,负责从经典方剂著作中准确提取方剂的组成和配伍关系。'
+ en: 'You are a professional traditional Chinese medicine scholar responsible for accurately extracting formula composition and compatibility from classical formula texts.'
+ rules_for_entities:
+ zh:
+ - '提取方剂名称作为实体'
+ - '提取方剂中的药材作为实体'
+ - '角色分类应遵循中医方剂学理论'
+ en:
+ - 'Extract formula names as entities'
+ - 'Extract herbs in formulas as entities'
+ - 'Role classification should follow TCM formula theory'
+ rules_for_relations:
+ zh:
+ - 'type为contains表示药材属于某方剂'
+ - 'type为antagonizes表示相恶相反的配伍禁忌'
+ - 'role字段应准确标注君臣佐使角色'
+ en:
+ - 'type contains indicates herb belongs to a formula'
+ - 'type antagonizes indicates incompatible compatibility'
+ - 'Role field should accurately indicate sovereign/minister/assistant/envoy role'
+
+identifiers:
+ entity_id: name
+ relation_id: '{formula}|{type}|{herb}'
+ relation_members: herb
+
+display:
+ entity_label: '{name}'
+ relation_label: '{role}'
diff --git a/hyperextract/templates/presets/tcm/herb_property.yaml b/hyperextract/templates/presets/tcm/herb_property.yaml
new file mode 100644
index 0000000..0aa623a
--- /dev/null
+++ b/hyperextract/templates/presets/tcm/herb_property.yaml
@@ -0,0 +1,74 @@
+language: [zh, en]
+
+name: herb_property
+type: model
+tags: [tcm, herb, traditional-chinese-medicine]
+
+description:
+ zh: '药材属性模型模板 - 从中药学教材中提取药材的基本属性信息'
+ en: 'Herb Property Model Template - Extract basic property information of medicinal herbs from traditional Chinese medicine textbooks'
+
+output:
+ description:
+ zh: '药材的基本属性'
+ en: 'Basic properties of medicinal herbs'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '药材名称'
+ en: 'Herb name'
+ required: true
+ - name: nature
+ type: str
+ description:
+ zh: '四气:寒/热/温/凉/平'
+ en: 'Four natures: cold/hot/warm/cool/neutral'
+ required: true
+ - name: flavor
+ type: str
+ description:
+ zh: '五味:酸/苦/甘/辛/咸'
+ en: 'Five flavors: sour/bitter/sweet/pungent/salty'
+ required: true
+ - name: meridian_entry
+ type: str
+ description:
+ zh: '归经'
+ en: 'Meridian entry'
+ required: true
+ - name: efficacy
+ type: str
+ description:
+ zh: '主要功效'
+ en: 'Main efficacy'
+ required: true
+ - name: toxicity
+ type: str
+ description:
+ zh: '毒性(无毒/小毒/有毒)'
+ en: 'Toxicity (non-toxic/slightly toxic/toxic)'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的中医药学家,负责从中药学教材中准确提取药材的基本属性信息。'
+ en: 'You are a professional traditional Chinese medicine scholar responsible for accurately extracting basic property information of medicinal herbs from TCM textbooks.'
+ rules:
+ zh:
+ - '药材名称应使用规范的中药学名称'
+ - '四气五味应按照中医药理论准确分类'
+ - '归经应标注药物作用的经络'
+ - '功效应概括主要治疗作用'
+ - '毒性信息应准确标注'
+ en:
+ - 'Herb names should use standard TCM nomenclature'
+ - 'Four natures and five flavors should be accurately classified according to TCM theory'
+ - 'Meridian entry should indicate the meridians the drug acts on'
+ - 'Efficacy should summarize the main therapeutic effects'
+ - 'Toxicity information should be accurately labeled'
+
+identifiers: {}
+
+display:
+ label: '{name}'
diff --git a/hyperextract/templates/presets/tcm/herb_relation.yaml b/hyperextract/templates/presets/tcm/herb_relation.yaml
new file mode 100644
index 0000000..fcfbc93
--- /dev/null
+++ b/hyperextract/templates/presets/tcm/herb_relation.yaml
@@ -0,0 +1,94 @@
+language: [zh, en]
+
+name: herb_relation
+type: graph
+tags: [tcm, herb, compatibility, interaction]
+
+description:
+ zh: '药材关联图模板 - 从方剂学文献中提取药材之间的配伍关系'
+ en: 'Herb Relation Graph Template - Extract herb compatibility relationships from formula literature'
+
+output:
+ description:
+ zh: '药材之间的配伍关系网络'
+ en: 'Herb compatibility relationship network'
+ entities:
+ description:
+ zh: '药材实体'
+ en: 'Herb entities'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '药材名称'
+ en: 'Herb name'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '药材说明'
+ en: 'Herb description'
+ required: false
+ relations:
+ description:
+ zh: '药材之间的配伍关系'
+ en: 'Herb compatibility relationships'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '药材A'
+ en: 'Herb A'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '药材B'
+ en: 'Herb B'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '配伍类型:相须/相使/相畏/相杀/相恶/相反'
+ en: 'Compatibility type: synergistic/mutual-enhancing/mutual-repressing/toxicity-reducing/antagonistic/incompatible'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '配伍说明'
+ en: 'Compatibility description'
+ required: false
+
+guideline:
+ target:
+ zh: '你是一位专业的中医药学家,负责从方剂学文献中准确提取药材之间的配伍关系。'
+ en: 'You are a professional traditional Chinese medicine scholar responsible for accurately extracting herb compatibility relationships from formula literature.'
+ rules_for_entities:
+ zh:
+ - '提取药材名称作为实体'
+ - '避免重复的药材实体'
+ - '描述应简洁概括药材特性'
+ en:
+ - 'Extract herb names as entities'
+ - 'Avoid duplicate herb entities'
+ - 'Description should concisely summarize herb characteristics'
+ rules_for_relations:
+ zh:
+ - '优先提取方剂学文献中明确描述的药材配伍关系'
+ - '避免重复的配伍关系记录'
+ - '仅提取文本中明确表达的配伍关系'
+ en:
+ - 'Prioritize explicitly described herb compatibility relationships from formula literature'
+ - 'Avoid duplicate compatibility relationship records'
+ - 'Only extract compatibility relationships explicitly expressed in the text'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/tcm/meridian_graph.yaml b/hyperextract/templates/presets/tcm/meridian_graph.yaml
new file mode 100644
index 0000000..8c3d2e6
--- /dev/null
+++ b/hyperextract/templates/presets/tcm/meridian_graph.yaml
@@ -0,0 +1,92 @@
+language: [zh, en]
+
+name: meridian_graph
+type: graph
+tags: [tcm, meridian, acupuncture, acupoint]
+
+description:
+ zh: '经络流注图模板 - 从针灸学教材中提取穴位和经脉的关系'
+ en: 'Meridian Flow Graph Template - Extract relationships between acupoints and meridians from acupuncture textbooks'
+
+output:
+ description:
+ zh: '穴位和经脉的关系网络'
+ en: 'Network of acupoint and meridian relationships'
+ entities:
+ description:
+ zh: '穴位和经脉'
+ en: 'Acupoints and meridians'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '穴位或经脉名称'
+ en: 'Acupoint or meridian name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '实体类型:穴位/经脉'
+ en: 'Entity type: acupoint/meridian'
+ required: true
+ - name: location
+ type: str
+ description:
+ zh: '位置描述'
+ en: 'Location description'
+ required: false
+ relations:
+ description:
+ zh: '穴位和经脉的归属和流注关系'
+ en: 'Belonging and flow relationships between acupoints and meridians'
+ fields:
+ - name: source
+ type: str
+ description:
+ zh: '源实体'
+ en: 'Source entity'
+ required: true
+ - name: target
+ type: str
+ description:
+ zh: '目标实体'
+ en: 'Target entity'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:located_on/flows_to'
+ en: 'Relation type: located_on/flows_to'
+ required: true
+
+guideline:
+ target:
+ zh: '你是一位专业的针灸学专家,负责从针灸学教材中准确提取穴位和经脉的关系。'
+ en: 'You are a professional acupuncture expert responsible for accurately extracting relationships between acupoints and meridians from acupuncture textbooks.'
+ rules_for_entities:
+ zh:
+ - '提取穴位名称作为实体'
+ - '提取经脉名称作为实体'
+ - '位置描述应准确标注解剖位置'
+ en:
+ - 'Extract acupoint names as entities'
+ - 'Extract meridian names as entities'
+ - 'Location descriptions should accurately indicate anatomical positions'
+ rules_for_relations:
+ zh:
+ - 'type为located_on表示穴位属于某经脉'
+ - 'type为flows_to表示经脉的流注顺序'
+ en:
+ - 'type located_on indicates acupoint belongs to a meridian'
+ - 'type flows_to indicates the flow sequence of meridians'
+
+identifiers:
+ entity_id: name
+ relation_id: '{source}|{type}|{target}'
+ relation_members:
+ source: source
+ target: target
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'
diff --git a/hyperextract/templates/presets/tcm/syndrome_reasoning.yaml b/hyperextract/templates/presets/tcm/syndrome_reasoning.yaml
new file mode 100644
index 0000000..f6a23db
--- /dev/null
+++ b/hyperextract/templates/presets/tcm/syndrome_reasoning.yaml
@@ -0,0 +1,106 @@
+language: [zh, en]
+
+name: syndrome_reasoning
+type: hypergraph
+tags: [tcm, syndrome, diagnosis, reasoning]
+
+description:
+ zh: '证候推理图模板 - 从医案和临床经验中提取症状、证候和治法的推理关系'
+ en: 'Syndrome Reasoning Graph Template - Extract symptom-syndrome-treatment reasoning relationships from medical cases and clinical experience'
+
+output:
+ description:
+ zh: '症状到证候到治法的推理链条'
+ en: 'Reasoning chain from symptoms to syndrome to treatment'
+ entities:
+ description:
+ zh: '症状、证候和治法要素'
+ en: 'Symptom, syndrome, and treatment elements'
+ fields:
+ - name: name
+ type: str
+ description:
+ zh: '要素名称'
+ en: 'Element name'
+ required: true
+ - name: type
+ type: str
+ description:
+ zh: '要素类型:症状/证候/治法/方剂'
+ en: 'Element type: symptom/syndrome/treatment/formula'
+ required: true
+ - name: description
+ type: str
+ description:
+ zh: '说明'
+ en: 'Description'
+ required: false
+ relations:
+ description:
+ zh: '症状、证候、治法之间的推理关系'
+ en: 'Reasoning relationships between symptoms, syndromes, and treatments'
+ fields:
+ - name: symptom
+ type: list
+ description:
+ zh: '症状表现'
+ en: 'Symptom manifestations'
+ required: false
+ - name: syndrome
+ type: str
+ description:
+ zh: '证候类型'
+ en: 'Syndrome type'
+ required: true
+ - name: principle
+ type: str
+ description:
+ zh: '治法原则'
+ en: 'Treatment principle'
+ required: true
+ - name: formula
+ type: str
+ description:
+ zh: '推荐方剂'
+ en: 'Recommended formula'
+ required: false
+ - name: type
+ type: str
+ description:
+ zh: '关系类型:leads_to/treated_by'
+ en: 'Relation type: leads_to/treated_by'
+ required: true
+
+guideline:
+ target:
+ zh: '你是一位专业的中医临床专家,负责从医案和临床经验中准确提取证候推理关系。'
+ en: 'You are a professional TCM clinical expert responsible for accurately extracting syndrome reasoning relationships from medical cases and clinical experience.'
+ rules_for_entities:
+ zh:
+ - '提取症状表现作为实体'
+ - '提取证候类型作为实体'
+ - '提取治法原则作为实体'
+ - '提取推荐方剂作为实体'
+ en:
+ - 'Extract symptom manifestations as entities'
+ - 'Extract syndrome types as entities'
+ - 'Extract treatment principles as entities'
+ - 'Extract recommended formulas as entities'
+ rules_for_relations:
+ zh:
+ - 'type为leads_to表示从症状到证候的推导'
+ - 'type为treated_by表示从证候到治法的对应'
+ - 'symptom字段应包含所有相关的症状'
+ en:
+ - 'type leads_to indicates derivation from symptoms to syndrome'
+ - 'type treated_by indicates correspondence from syndrome to treatment'
+ - 'Symptom field should include all related symptoms'
+
+identifiers:
+ entity_id: name
+ relation_id: '{syndrome}|{type}|{principle}'
+ relation_members: symptom
+
+display:
+ entity_label: '{name}'
+ relation_label: '{type}'