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: 添加教育领域的先修图谱、跨学科超图和学术词汇模板,支持多语言

yifanfeng97d7ecf74b816d
diff --git a/hyperextract/templates/en/education/__init__.py b/hyperextract/templates/en/education/__init__.py
index e69de29..7aee758 100644
--- a/hyperextract/templates/en/education/__init__.py
+++ b/hyperextract/templates/en/education/__init__.py
@@ -0,0 +1,9 @@
+from .prerequisite_map import PrerequisiteMapGraph
+from .academic_lexicon import AcademicLexiconSet
+from .interdisciplinary_nexus import InterdisciplinaryNexusHypergraph
+
+__all__ = [
+ "PrerequisiteMapGraph",
+ "AcademicLexiconSet",
+ "InterdisciplinaryNexusHypergraph",
+]
diff --git a/hyperextract/templates/en/education/academic_lexicon.py b/hyperextract/templates/en/education/academic_lexicon.py
new file mode 100644
index 0000000..c016360
--- /dev/null
+++ b/hyperextract/templates/en/education/academic_lexicon.py
@@ -0,0 +1,99 @@
+from typing import Optional, List, Any
+from pydantic import BaseModel, Field
+from langchain_core.language_models.chat_models import BaseChatModel
+from langchain_core.embeddings import Embeddings
+from hyperextract.types import AutoSet
+
+# ==============================================================================
+# 1. Schema Definitions
+# ==============================================================================
+
+class AcademicTerm(BaseModel):
+ """
+ A standardized academic term, concept, or definition aggregated from multiple sources.
+ """
+ term: str = Field(
+ description="The primary name of the academic term/concept. Acts as the unique identifier."
+ )
+ domain: str = Field(
+ description="Field of study or academic domain (e.g., 'Thermodynamics', 'Linguistics')."
+ )
+ definition: str = Field(
+ description="A comprehensive, synthesized definition of the term."
+ )
+ synonyms: List[str] = Field(
+ default_factory=list,
+ description="Alternative names or closely related terms."
+ )
+ key_expression: Optional[str] = Field(
+ None, description="A formula, notation, or iconic expression associated with this term."
+ )
+
+# ==============================================================================
+# 2. Prompts
+# ==============================================================================
+
+ACADEMIC_LEXICON_PROMPT = (
+ "You are a Lexicographer and Academic Encyclopedia Editor. Your goal is to extract standardize academic definitions.\n\n"
+ "Extraction Rules:\n"
+ "- Identify unique academic terms and their highly precise definitions.\n"
+ "- Search for alternative names or synonyms within the text.\n"
+ "- If multiple descriptions of the same term exist, prioritize the one that is most mathematically or logically rigorous."
+)
+
+# ==============================================================================
+# 3. Template Class
+# ==============================================================================
+
+class AcademicLexiconSet(AutoSet[AcademicTerm]):
+ """
+ Educational template for building a deduplicated, synthesized dictionary of academic terms.
+ Leverages AutoSet to merge conflicting or complementary definitions from multiple text sources.
+
+ Example Usage:
+ >>> from langchain_openai import ChatOpenAI, OpenAIEmbeddings
+ >>> from hyperextract.templates.en.education.academic_lexicon import AcademicLexiconSet
+ >>> lexicon = AcademicLexiconSet(llm_client, embedder)
+ >>> lexicon.feed_text("Source 1: Photosynthesis is a process used by plants...")
+ >>> lexicon.feed_text("Source 2: Photosynthesis converts light energy into chemical energy...")
+ >>> lexicon.extract() # Will merge definitions of "Photosynthesis"
+ """
+
+ def __init__(
+ self,
+ llm_client: BaseChatModel,
+ embedder: Embeddings,
+ *,
+ extraction_mode: str = "one_stage", # For AutoSet, this is standard
+ chunk_size: int = 2048,
+ chunk_overlap: int = 256,
+ max_workers: int = 10,
+ verbose: bool = False,
+ **kwargs: Any,
+ ):
+ """
+ Initialize the Academic Lexicon Template.
+
+ Args:
+ llm_client (BaseChatModel): The LLM used for extraction.
+ embedder (Embeddings): The embedding model for entity deduplication and merging.
+ extraction_mode (str): Mode of extraction.
+ chunk_size (int): Size of text chunks for processing large documents.
+ chunk_overlap (int): Overlap between chunks.
+ max_workers (int): Parallel processing workers.
+ verbose (bool): Enable verbose logging.
+ **kwargs: Additional parameters for AutoSet.
+ """
+ super().__init__(
+ item_schema=AcademicTerm,
+ key_extractor=lambda x: x.term.strip().lower(),
+ llm_client=llm_client,
+ embedder=embedder,
+ extraction_mode=extraction_mode,
+ chunk_size=chunk_size,
+ chunk_overlap=chunk_overlap,
+ max_workers=max_workers,
+ verbose=verbose,
+ prompt=ACADEMIC_LEXICON_PROMPT,
+ **kwargs,
+ )
diff --git a/hyperextract/templates/en/education/interdisciplinary_nexus.py b/hyperextract/templates/en/education/interdisciplinary_nexus.py
new file mode 100644
index 0000000..714d87e
--- /dev/null
+++ b/hyperextract/templates/en/education/interdisciplinary_nexus.py
@@ -0,0 +1,118 @@
+from typing import List, Any
+from pydantic import BaseModel, Field
+from langchain_core.language_models.chat_models import BaseChatModel
+from langchain_core.embeddings import Embeddings
+from hyperextract.types import AutoHypergraph
+
+# ==============================================================================
+# 1. Schema Definitions
+# ==============================================================================
+
+class ConceptNode(BaseModel):
+ """
+ A fundamental academic concept that might belong to multiple disciplines.
+ """
+ name: str = Field(description="Standard name of the concept.")
+ primary_field: str = Field(description="The primary discipline this concept originates from.")
+
+class InterdisciplinaryEdge(BaseModel):
+ """
+ A hyperedge representing a connection between multiple concepts across different fields.
+ """
+ participants: List[str] = Field(
+ description="List of nodes (concept names) participating in this interdisciplinary nexus."
+ )
+ unifying_theme: str = Field(
+ description="The overarching theme or problem that connects these concepts (e.g., 'Complexity Theory', 'Universal Design')."
+ )
+ connectivity_insight: str = Field(
+ description="The scientific or pedagogical value of connecting these specific dots."
+ )
+
+# ==============================================================================
+# 2. Prompts
+# ==============================================================================
+
+NEXUS_CONSOLIDATED_PROMPT = (
+ "You are a Polymath and Interdisciplinary Researcher. Identify deep connections between concepts across different academic fields.\n\n"
+ "Extraction Rules:\n"
+ "- Identify core concepts as Nodes.\n"
+ "- Create Hyperedges that group multiple concepts under a unifying interdisciplinary theme.\n"
+ "- Focus on finding surprising but scientifically valid bridges (e.g., connecting Biology and Economics through Game Theory)."
+)
+
+NEXUS_NODE_PROMPT = (
+ "Identify fundamental academic concepts that carry cross-disciplinary potential.\n\n"
+ "Rules:\n"
+ "- Extract nouns/concepts that are central to their respective fields.\n"
+ "- Identify the field of origin."
+)
+
+NEXUS_EDGE_PROMPT = (
+ "Identify the 'Nexuses' or Hyperedges that unify a group of concepts.\n\n"
+ "Rules:\n"
+ "- A hyperedge must involve more than two nodes whenever possible.\n"
+ "- Describe the unifying theme and the insight gained by this connection."
+)
+
+# ==============================================================================
+# 3. Template Class
+# ==============================================================================
+
+class InterdisciplinaryNexusHypergraph(AutoHypergraph[ConceptNode, InterdisciplinaryEdge]):
+ """
+ Educational template for mapping non-binary, multi-point academic connections.
+ Ideal for research synthesis, curriculum cross-linking, and holistic knowledge management.
+
+ Example Usage:
+ >>> nexus = InterdisciplinaryNexusHypergraph(llm, embedder)
+ >>> nexus.feed_text("The concept of Entropy applies to Heat (Physics) and Information (Information Theory).")
+ >>> nexus.extract()
+ >>> # Result will contain a Hyperedge connecting 'Heat' and 'Information' via 'Entropy'.
+ """
+
+ def __init__(
+ self,
+ llm_client: BaseChatModel,
+ embedder: Embeddings,
+ *,
+ extraction_mode: str = "one_stage",
+ chunk_size: int = 2048,
+ chunk_overlap: int = 256,
+ max_workers: int = 10,
+ verbose: bool = False,
+ **kwargs: Any,
+ ):
+ """
+ Initialize the Interdisciplinary Nexus Template.
+
+ Args:
+ llm_client (BaseChatModel): The LLM used for extraction.
+ embedder (Embeddings): The embedding model for deduplication.
+ extraction_mode (str): 'one_stage' or 'two_stage'.
+ chunk_size (int): Size of chunks.
+ chunk_overlap (int): Overlap.
+ max_workers (int): Workers.
+ verbose (bool): Logs.
+ **kwargs: Args for AutoHypergraph.
+ """
+ super().__init__(
+ node_schema=ConceptNode,
+ edge_schema=InterdisciplinaryEdge,
+ node_key_extractor=lambda x: x.name.strip().lower(),
+ edge_key_extractor=lambda x: (
+ f"{x.unifying_theme.strip().lower()}_linking_{sorted(x.participants)}"
+ ),
+ nodes_in_edge_extractor=lambda x: [p.strip().lower() for p in x.participants],
+ llm_client=llm_client,
+ embedder=embedder,
+ extraction_mode=extraction_mode,
+ chunk_size=chunk_size,
+ chunk_overlap=chunk_overlap,
+ max_workers=max_workers,
+ verbose=verbose,
+ prompt=NEXUS_CONSOLIDATED_PROMPT,
+ prompt_for_node_extraction=NEXUS_NODE_PROMPT,
+ prompt_for_edge_extraction=NEXUS_EDGE_PROMPT,
+ **kwargs,
+ )
diff --git a/hyperextract/templates/en/education/prerequisite_map.py b/hyperextract/templates/en/education/prerequisite_map.py
new file mode 100644
index 0000000..2e1da13
--- /dev/null
+++ b/hyperextract/templates/en/education/prerequisite_map.py
@@ -0,0 +1,134 @@
+from typing import Optional, Any
+from pydantic import BaseModel, Field
+from langchain_core.language_models.chat_models import BaseChatModel
+from langchain_core.embeddings import Embeddings
+from hyperextract.types import AutoGraph
+
+# ==============================================================================
+# 1. Schema Definitions
+# ==============================================================================
+
+class KnowledgePointNode(BaseModel):
+ """
+ A discrete unit of knowledge, concept, or skill within a specific curriculum.
+ """
+ name: str = Field(
+ description="Standardized name of the knowledge point or concept."
+ )
+ level: str = Field(
+ description="Difficulty level: 'Primary', 'Intermediate', 'Advanced', 'Expert'."
+ )
+ concept_type: str = Field(
+ description="Pedagogical type: 'Definition', 'Theorem', 'Formula', 'Method', 'Application'."
+ )
+ description: Optional[str] = Field(
+ None, description="A brief pedagogical summary of what this concept entails."
+ )
+
+class PrerequisiteRelation(BaseModel):
+ """
+ Directional dependency indicating that one concept must be learned before another.
+ """
+ source: str = Field(description="The prerequisite knowledge point (must be learned first).")
+ target: str = Field(description="The successor knowledge point (learned after).")
+ dependency_type: str = Field(
+ "Prerequisite",
+ description="Nature of link: 'Prerequisite' (strict), 'Corequisite' (together), 'Recommended' (helpful)."
+ )
+ weight: float = Field(
+ 1.0, description="Strength of dependency from 0.0 to 1.0."
+ )
+
+# ==============================================================================
+# 2. Prompts
+# ==============================================================================
+
+EDUCATION_CONSOLIDATED_PROMPT = (
+ "You are a Senior Curriculum Architect and Educational Psychologist. Extract the logical structure of knowledge points from the text.\n\n"
+ "Rules:\n"
+ "- Identify atomic concepts that constitute the learning path.\n"
+ "- Establish clear directional dependencies (prerequisites).\n"
+ "- Ensure that the resulting graph reflects a sound pedagogical sequence."
+)
+
+EDUCATION_NODE_PROMPT = (
+ "You are an Educational Psychologist specializing in knowledge decomposition. Your task is to identify all discrete Knowledge Points (Nodes).\n\n"
+ "Extraction Rules:\n"
+ "- Focus on nouns or noun phrases that represent specific academic concepts, formulas, or methods.\n"
+ "- Assign a difficulty level and concept type based on the context.\n"
+ "- DO NOT identify relationships or dependencies at this stage."
+)
+
+EDUCATION_EDGE_PROMPT = (
+ "You are a Curriculum Designer. Given the identified knowledge points, map the Prerequisite Relations (Edges).\n\n"
+ "Extraction Rules:\n"
+ "- Determine if Concept A is a necessary precursor to Concept B.\n"
+ "- Classify the dependency as Prerequisite, Corequisite, or Recommended.\n"
+ "- Only create relationships between concepts identified in the provided list."
+)
+
+# ==============================================================================
+# 3. Template Class
+# ==============================================================================
+
+class PrerequisiteMapGraph(AutoGraph[KnowledgePointNode, PrerequisiteRelation]):
+ """
+ Educational template for building Directed Acyclic Graphs (DAG) of knowledge dependencies.
+ Useful for syllabus analysis, learning path generation, and student assessment planning.
+
+ Example Usage:
+ >>> from langchain_openai import ChatOpenAI, OpenAIEmbeddings
+ >>> from hyperextract.templates.en.education.prerequisite_map import PrerequisiteMapGraph
+ >>> llm = ChatOpenAI(model="gpt-4o")
+ >>> embedder = OpenAIEmbeddings()
+ >>> graph = PrerequisiteMapGraph(llm, embedder)
+ >>> graph.feed_text("To understand Calculus, one must first master Algebra and Functions.")
+ >>> graph.extract()
+ >>> print(graph.nodes)
+ """
+
+ def __init__(
+ self,
+ llm_client: BaseChatModel,
+ embedder: Embeddings,
+ *,
+ extraction_mode: str = "one_stage",
+ chunk_size: int = 2048,
+ chunk_overlap: int = 256,
+ max_workers: int = 10,
+ verbose: bool = False,
+ **kwargs: Any,
+ ):
+ """
+ Initialize the Prerequisite Map Template.
+
+ Args:
+ llm_client (BaseChatModel): The LLM used for extraction.
+ embedder (Embeddings): The embedding model for node/edge deduplication.
+ extraction_mode (str): 'one_stage' for consolidated extraction, 'two_stage' for separate node/edge passes.
+ chunk_size (int): Size of text chunks for processing large documents.
+ chunk_overlap (int): Overlap between chunks to preserve context.
+ max_workers (int): Number of parallel workers for processing.
+ verbose (bool): Whether to print detailed logs during execution.
+ **kwargs: Additional parameters passed to the AutoGraph base class.
+ """
+ super().__init__(
+ node_schema=KnowledgePointNode,
+ edge_schema=PrerequisiteRelation,
+ node_key_extractor=lambda x: x.name.strip().lower(),
+ edge_key_extractor=lambda x: (
+ f"{x.source.strip().lower()}->{x.target.strip().lower()}"
+ ),
+ nodes_in_edge_extractor=lambda x: (x.source.strip().lower(), x.target.strip().lower()),
+ llm_client=llm_client,
+ embedder=embedder,
+ extraction_mode=extraction_mode,
+ chunk_size=chunk_size,
+ chunk_overlap=chunk_overlap,
+ max_workers=max_workers,
+ verbose=verbose,
+ prompt=EDUCATION_CONSOLIDATED_PROMPT,
+ prompt_for_node_extraction=EDUCATION_NODE_PROMPT,
+ prompt_for_edge_extraction=EDUCATION_EDGE_PROMPT,
+ **kwargs,
+ )
diff --git a/hyperextract/templates/zh/education/__init__.py b/hyperextract/templates/zh/education/__init__.py
index e69de29..7aee758 100644
--- a/hyperextract/templates/zh/education/__init__.py
+++ b/hyperextract/templates/zh/education/__init__.py
@@ -0,0 +1,9 @@
+from .prerequisite_map import PrerequisiteMapGraph
+from .academic_lexicon import AcademicLexiconSet
+from .interdisciplinary_nexus import InterdisciplinaryNexusHypergraph
+
+__all__ = [
+ "PrerequisiteMapGraph",
+ "AcademicLexiconSet",
+ "InterdisciplinaryNexusHypergraph",
+]
diff --git a/hyperextract/templates/zh/education/academic_lexicon.py b/hyperextract/templates/zh/education/academic_lexicon.py
new file mode 100644
index 0000000..68b80cd
--- /dev/null
+++ b/hyperextract/templates/zh/education/academic_lexicon.py
@@ -0,0 +1,97 @@
+from typing import Optional, List, Any
+from pydantic import BaseModel, Field
+from langchain_core.language_models.chat_models import BaseChatModel
+from langchain_core.embeddings import Embeddings
+from hyperextract.types import AutoSet
+
+# ==============================================================================
+# 1. Schema 定义
+# ==============================================================================
+
+class AcademicTerm(BaseModel):
+ """
+ 从多个来源聚合的标准学术词条、概念或定义。
+ """
+ term: str = Field(
+ description="学术词条/概念的主名称。作为唯一标识符。"
+ )
+ domain: str = Field(
+ description="研究领域或学科(如:'热力学'、'语言学')。"
+ )
+ definition: str = Field(
+ description="综合多个来源后形成的权威、详细的定义描述。"
+ )
+ synonyms: List[str] = Field(
+ default_factory=list,
+ description="同义词或高度相关的别称列表。"
+ )
+ key_expression: Optional[str] = Field(
+ None, description="与该词条相关的核心公式、符号或标志性表达。"
+ )
+
+# ==============================================================================
+# 2. Prompts 提示词
+# ==============================================================================
+
+ACADEMIC_LEXICON_PROMPT = (
+ "你是一位学术百科全书编辑和词典编纂者。你的目标是提取并标准化学术定义。\n\n"
+ "提取规则:\n"
+ "- 识别唯一的学术术语及其高度精确的定义。\n"
+ "- 寻找文本中出现的同义词、简称或别称。\n"
+ "- 如果存在对同一术语的多个描述,优先选择数学上或逻辑上更严谨的描述,并尝试进行信息互补。"
+)
+
+# ==============================================================================
+# 3. 模板类
+# ==============================================================================
+
+class AcademicLexiconSet(AutoSet[AcademicTerm]):
+ """
+ 用于构建去重且合成的学术名词百科集的教育模版。
+ 利用 AutoSet 的合并策略,将来自不同教材或文献的定义融合成唯一的、信息最全的条目。
+
+ 用法示例:
+ >>> lexicon = AcademicLexiconSet(llm_client, embedder)
+ >>> lexicon.feed_text("来源A:光合作用是植物利用光能...")
+ >>> lexicon.feed_text("来源B:光合作用涉及叶绿素吸收光子...")
+ >>> lexicon.extract() # 自动通过“项合并”功能将关于“光合作用”的不同侧重定义合并
+ """
+
+ def __init__(
+ self,
+ llm_client: BaseChatModel,
+ embedder: Embeddings,
+ *,
+ extraction_mode: str = "one_stage",
+ chunk_size: int = 2048,
+ chunk_overlap: int = 256,
+ max_workers: int = 10,
+ verbose: bool = False,
+ **kwargs: Any,
+ ):
+ """
+ 初始化学术名词百科模板。
+
+ 参数描述:
+ llm_client (BaseChatModel): 用于提取的语言模型客户端。
+ embedder (Embeddings): 用于实体合并和去重的嵌入模型(AutoSet 的核心)。
+ extraction_mode (str): 提取模式策略。
+ chunk_size (int): 文本分片大小。
+ chunk_overlap (int): 分片重叠大小。
+ max_workers (int): 并行处理进程数。
+ verbose (bool): 是否开启详细日志。
+ **kwargs: 透传给 AutoSet 基类的其他参数(如自定义 merge_strategy)。
+ """
+ super().__init__(
+ item_schema=AcademicTerm,
+ key_extractor=lambda x: x.term.strip().lower(),
+ llm_client=llm_client,
+ embedder=embedder,
+ extraction_mode=extraction_mode,
+ chunk_size=chunk_size,
+ chunk_overlap=chunk_overlap,
+ max_workers=max_workers,
+ verbose=verbose,
+ prompt=ACADEMIC_LEXICON_PROMPT,
+ **kwargs,
+ )
diff --git a/hyperextract/templates/zh/education/interdisciplinary_nexus.py b/hyperextract/templates/zh/education/interdisciplinary_nexus.py
new file mode 100644
index 0000000..cd2086f
--- /dev/null
+++ b/hyperextract/templates/zh/education/interdisciplinary_nexus.py
@@ -0,0 +1,118 @@
+from typing import List, Any
+from pydantic import BaseModel, Field
+from langchain_core.language_models.chat_models import BaseChatModel
+from langchain_core.embeddings import Embeddings
+from hyperextract.types import AutoHypergraph
+
+# ==============================================================================
+# 1. Schema 定义
+# ==============================================================================
+
+class ConceptNode(BaseModel):
+ """
+ 可能属于多个学科的基础学术概念。
+ """
+ name: str = Field(description="概念的标准名称。")
+ primary_field: str = Field(description="该概念原生所属的主要学科。")
+
+class InterdisciplinaryEdge(BaseModel):
+ """
+ 表示不同领域多个概念之间复杂联系的超边。
+ """
+ participants: List[str] = Field(
+ description="参与该跨学科连接的所有节点名称列表。"
+ )
+ unifying_theme: str = Field(
+ description="将这些概念连接在一起的跨学科主题(例如:'复杂性理论'、'通用设计')。"
+ )
+ connectivity_insight: str = Field(
+ description="将这些特定知识点连接起来所产生的科学或教学价值。"
+ )
+
+# ==============================================================================
+# 2. Prompts 提示词
+# ==============================================================================
+
+NEXUS_CONSOLIDATED_PROMPT = (
+ "你是一位通才和跨学科研究专家。请识别不同学术领域概念之间的深层联系。\n\n"
+ "提取规则:\n"
+ "- 识别核心概念作为节点。\n"
+ "- 创建超边,将多个概念归纳在一个统一的跨学科主题下。\n"
+ "- 专注于发现令人惊讶但科学有效的桥梁连接(例如:通过博弈论连接生物学和经济学)。"
+)
+
+NEXUS_NODE_PROMPT = (
+ "识别具有跨学科潜力的基础学术概念。\n\n"
+ "提取规则:\n"
+ "- 提取对各自领域至关重要的名词/概念。\n"
+ "- 标注其学科来源。"
+)
+
+NEXUS_EDGE_PROMPT = (
+ "识别连接一组概念的“纽带”或称之为超边。\n\n"
+ "提取规则:\n"
+ "- 只要可能,一个超边应涉及两个以上的节点。\n"
+ "- 描述统一这些概念的主题以及这种连接提供的洞察力。"
+)
+
+# ==============================================================================
+# 3. 模板类
+# ==============================================================================
+
+class InterdisciplinaryNexusHypergraph(AutoHypergraph[ConceptNode, InterdisciplinaryEdge]):
+ """
+ 用于映射多点、非二元学术联系的教育领域模板。
+ 理想用于研究综述、课程跨学科设计以及跨学科知识管理。
+
+ 用法示例:
+ >>> nexus = InterdisciplinaryNexusHypergraph(llm, embedder)
+ >>> nexus.feed_text("‘熵’这一概念同时适用于热力学中的‘热’和信息论中的‘信息’。")
+ >>> nexus.extract()
+ >>> # 结果将包含一个通过“熵”这一主题连接“热”与“信息”的超边。
+ """
+
+ def __init__(
+ self,
+ llm_client: BaseChatModel,
+ embedder: Embeddings,
+ *,
+ extraction_mode: str = "one_stage",
+ chunk_size: int = 2048,
+ chunk_overlap: int = 256,
+ max_workers: int = 10,
+ verbose: bool = False,
+ **kwargs: Any,
+ ):
+ """
+ 初始化跨学科纽带超图模板。
+
+ 参数描述:
+ llm_client (BaseChatModel): 用于提取的语言模型客户端。
+ embedder (Embeddings): 用于节点去重的嵌入模型。
+ extraction_mode (str): 提取模式。
+ chunk_size (int): 文本分片大小。
+ chunk_overlap (int): 分片重叠大小。
+ max_workers (int): 并行处理进程数。
+ verbose (bool): 是否打印日志。
+ **kwargs: 透传给 AutoHypergraph 基类的其他参数(如一致性检查开关)。
+ """
+ super().__init__(
+ node_schema=ConceptNode,
+ edge_schema=InterdisciplinaryEdge,
+ node_key_extractor=lambda x: x.name.strip().lower(),
+ edge_key_extractor=lambda x: (
+ f"{x.unifying_theme.strip().lower()}_linking_{sorted(x.participants)}"
+ ),
+ nodes_in_edge_extractor=lambda x: [p.strip().lower() for p in x.participants],
+ llm_client=llm_client,
+ embedder=embedder,
+ extraction_mode=extraction_mode,
+ chunk_size=chunk_size,
+ chunk_overlap=chunk_overlap,
+ max_workers=max_workers,
+ verbose=verbose,
+ prompt=NEXUS_CONSOLIDATED_PROMPT,
+ prompt_for_node_extraction=NEXUS_NODE_PROMPT,
+ prompt_for_edge_extraction=NEXUS_EDGE_PROMPT,
+ **kwargs,
+ )
diff --git a/hyperextract/templates/zh/education/prerequisite_map.py b/hyperextract/templates/zh/education/prerequisite_map.py
new file mode 100644
index 0000000..19fca73
--- /dev/null
+++ b/hyperextract/templates/zh/education/prerequisite_map.py
@@ -0,0 +1,132 @@
+from typing import Optional, Any
+from pydantic import BaseModel, Field
+from langchain_core.language_models.chat_models import BaseChatModel
+from langchain_core.embeddings import Embeddings
+from hyperextract.types import AutoGraph
+
+# ==============================================================================
+# 1. Schema 定义
+# ==============================================================================
+
+class KnowledgePointNode(BaseModel):
+ """
+ 学科大纲中的原子知识单位、概念或技能。
+ """
+ name: str = Field(
+ description="知识点或概念的标准名称。"
+ )
+ level: str = Field(
+ description="难度等级:'入门'、'中级'、'高级'、'专家'。"
+ )
+ concept_type: str = Field(
+ description="教学类型:'定义'、'定理'、'公式'、'方法'、'应用'。"
+ )
+ description: Optional[str] = Field(
+ None, description="该概念的核心教学内涵简述。"
+ )
+
+class PrerequisiteRelation(BaseModel):
+ """
+ 有向依赖关系,表示一个概念必须在另一个概念之前学习。
+ """
+ source: str = Field(description="先修知识点(必须先学)。")
+ target: str = Field(description="后续知识点(后学)。")
+ dependency_type: str = Field(
+ "Prerequisite",
+ description="依赖性质:'必要' (必须先学), '同步' (建议同时学), '辅助' (有助于学习)。"
+ )
+ weight: float = Field(
+ 1.0, description="依赖强度,范围从 0.0 到 1.0。"
+ )
+
+# ==============================================================================
+# 2. Prompts 提示词
+# ==============================================================================
+
+EDUCATION_CONSOLIDATED_PROMPT = (
+ "你是一位资深课程架构师和教育心理学家。请从文本中提取知识点的逻辑结构。\n\n"
+ "提取规则:\n"
+ "- 识别构成学习路径的原子化概念。\n"
+ "- 建立清晰的有向依赖关系(先修要求)。\n"
+ "- 确保生成的图谱符合科学的教学演进逻辑。"
+)
+
+EDUCATION_NODE_PROMPT = (
+ "你是一位擅长知识拆解的教育心理学家。你的任务是识别所有离散的知识点(节点)。\n\n"
+ "提取规则:\n"
+ "- 关注代表特定学术概念、公式或方法的名词或名词短语。\n"
+ "- 根据上下文分配难度等级和概念类型。\n"
+ "- 此阶段不要识别关系或依赖。"
+)
+
+EDUCATION_EDGE_PROMPT = (
+ "你是一位课程设计师。根据已识别的知识点,映射先修关系(边)。\n\n"
+ "提取规则:\n"
+ "- 确定概念 A 是否是学习概念 B 的必要前提。\n"
+ "- 将依赖关系分类为必要、同步或辅助。\n"
+ "- 仅在提供的概念列表之间建立关系。"
+)
+
+# ==============================================================================
+# 3. 模板类
+# ==============================================================================
+
+class PrerequisiteMapGraph(AutoGraph[KnowledgePointNode, PrerequisiteRelation]):
+ """
+ 用于构建知识依赖有向无环图 (DAG) 的教育领域模板。
+ 适用于教学大纲分析、自动化学习路径生成和学生测评规划。
+
+ 用法示例:
+ >>> from langchain_openai import ChatOpenAI, OpenAIEmbeddings
+ >>> from hyperextract.templates.zh.education.prerequisite_map import PrerequisiteMapGraph
+ >>> graph = PrerequisiteMapGraph(llm, embedder)
+ >>> graph.feed_text("在理解微积分之前,必须先掌握代数和函数。")
+ >>> graph.extract()
+ >>> print(graph.nodes) # 输出包含代数、函数、微积分及其依赖关系的节点
+ """
+
+ def __init__(
+ self,
+ llm_client: BaseChatModel,
+ embedder: Embeddings,
+ *,
+ extraction_mode: str = "one_stage",
+ chunk_size: int = 2048,
+ chunk_overlap: int = 256,
+ max_workers: int = 10,
+ verbose: bool = False,
+ **kwargs: Any,
+ ):
+ """
+ 初始化先修图谱模板。
+
+ 参数描述:
+ llm_client (BaseChatModel): 用于提取的语言模型客户端。
+ embedder (Embeddings): 用于节点/边去重和合并的嵌入模型。
+ extraction_mode (str): 提取模式。'one_stage' 为单阶段合并提取, 'two_stage' 为节点和边分开提取。
+ chunk_size (int): 处理大文档时的文本切片大小。
+ chunk_overlap (int): 切片之间的重叠长度以保持上下文完整性。
+ max_workers (int): 并行处理的最大进程数。
+ verbose (bool): 是否在执行过程中打印详细日志。
+ **kwargs: 透传给 AutoGraph 基类的其他参数。
+ """
+ super().__init__(
+ node_schema=KnowledgePointNode,
+ edge_schema=PrerequisiteRelation,
+ node_key_extractor=lambda x: x.name.strip().lower(),
+ edge_key_extractor=lambda x: (
+ f"{x.source.strip().lower()}->{x.target.strip().lower()}"
+ ),
+ nodes_in_edge_extractor=lambda x: (x.source.strip().lower(), x.target.strip().lower()),
+ llm_client=llm_client,
+ embedder=embedder,
+ extraction_mode=extraction_mode,
+ chunk_size=chunk_size,
+ chunk_overlap=chunk_overlap,
+ max_workers=max_workers,
+ verbose=verbose,
+ prompt=EDUCATION_CONSOLIDATED_PROMPT,
+ prompt_for_node_extraction=EDUCATION_NODE_PROMPT,
+ prompt_for_edge_extraction=EDUCATION_EDGE_PROMPT,
+ **kwargs,
+ )