Graph-Native Infrastructure for Context and Accountable AI Systems
Python agent-memoryaiai-governanceai-infrastructureartificial-intelligencecontext-engineeringcontext-graphsdata-engineeringdecision-intelligencedeveloper-toolsexplainable-aigenerative-aigraph-ragknowledge-graphllmontologyprovenancepythonreasoningsemantic-search
docs(learning-more): fix broken pipeline and dedup snippets (#1511)
Zohaib Hassnain82fd1b8d885d
diff --git a/docs/learning-more.md b/docs/learning-more.mdindex 6f74d88c..ab49bad3 100644--- a/docs/learning-more.md+++ b/docs/learning-more.md@@ -64,7 +64,7 @@ Whether you're running your first pipeline or deploying Semantica in production,[Temporal Graphs notebook](https://github.com/semantica-agi/semantica/blob/main/cookbook/advanced/10_Temporal_Knowledge_Graphs.ipynb): `valid_from`/`valid_until`, Allen interval algebra, point-in-time queries.</Step><Step title="Ontology-driven knowledge bases">- [Ontology notebook](https://github.com/semantica-agi/semantica/blob/main/cookbook/introduction/14_Ontology.ipynb): auto-generation, SHACL validation, Ontology Hub (v0.5.0).+ [Ontology notebook](https://github.com/semantica-agi/semantica/blob/main/cookbook/introduction/14_Ontology.ipynb): auto-generation, SHACL validation, Ontology Hub.</Step><Step title="Advanced visualization">[Complete Visualization Suite notebook](https://github.com/semantica-agi/semantica/blob/main/cookbook/advanced/03_Complete_Visualization_Suite.ipynb): UMAP, t-SNE, community layouts, embedding projections.@@ -86,10 +86,10 @@ All settings can be overridden with environment variables: no code changes neede| OpenAI API Key | `OPENAI_API_KEY` | `None` || Groq API Key | `GROQ_API_KEY` | `None` || Anthropic API Key | `ANTHROPIC_API_KEY` | `None` |-| Embedding Provider | `SEMANTICA_EMBEDDING_PROVIDER` | `"openai"` |-| Graph Backend | `SEMANTICA_GRAPH_BACKEND` | `"networkx"` |-| Log Level | `SEMANTICA_LOG_LEVEL` | `"INFO"` |-| Log Format | `SEMANTICA_LOG_FORMAT` | `"text"` |+| Graph Store Backend | `GRAPH_STORE_DEFAULT_BACKEND` | `"neo4j"` |+| Vector Store Backend | `VECTOR_STORE_DEFAULT_BACKEND` | `"faiss"` |+| Server Host | `SEMANTICA_HOST` | `"127.0.0.1"` |+| Server API Key | `SEMANTICA_API_KEY` | `None` |## Troubleshooting@@ -146,10 +146,15 @@ Also reduce batch sizes and enable streaming ingestion for large corpora.Enable parallel execution and GPU acceleration:```python-from semantica.pipeline import Pipeline--pipeline = Pipeline(workers=8, batch_size=32)-pipeline.run(sources)+from semantica.pipeline import ParallelismManager, Task++# Run pipeline tasks concurrently across worker threads+manager = ParallelismManager(max_workers=8)+tasks = [+ Task("task_1", lambda: "process part 1"),+ Task("task_2", lambda: "process part 2"),+]+results = manager.execute_parallel(tasks)``````bash@@ -160,19 +165,19 @@ pip install "semantica[gpu]" # CUDA-backed embeddings<Accordion title="Windows [all] installation fails" icon="windows">-Fixed in **v0.5.0**. Upgrade:+Upgrade to the latest release:```bashpip install --upgrade semantica```-Or install extras individually: `pip install "semantica[core]"`, then add `[llm-openai]`, `[gpu]`, etc. as needed.+Or install extras individually: `pip install semantica`, then add `[llm-openai]`, `[gpu]`, etc. as needed.</Accordion><Accordion title="cp1252 encoding crash on Windows" icon="windows">-Fixed in **v0.5.0**. For earlier versions, set the encoding environment variable:+Set the encoding environment variable:```bashset PYTHONIOENCODING=utf-8@@ -202,27 +207,44 @@ Use NetworkX for local development and prototyping. Switch to a persistent backe<Accordion title="Batch processing for large corpora" icon="layer-group">-Process documents in batches rather than one at a time. Configure `chunk_size` based on available RAM: a good starting point is 1,000 documents per batch on a 16 GB machine.+Process documents in batches rather than one at a time. Split large texts into chunks and extract entities in batches:```python-from semantica.pipeline import Pipeline+from semantica.split import TextSplitter+from semantica.semantic_extract import NERExtractor++document_text = "Acme Corp announced record revenue in Seattle. CEO Jane Doe presented results."+splitter = TextSplitter(chunk_size=1000, chunk_overlap=100)+chunks = splitter.split(document_text)-pipeline = Pipeline(workers=8, batch_size=32)-pipeline.run(sources)+extractor = NERExtractor()+batch_entities = extractor.extract_entities_batch([c.text for c in chunks])```</Accordion><Accordion title="Deduplication v2: up to 7× faster" icon="bolt">-If deduplication is a bottleneck, switch from v1 strategies to the v2 engine:+If deduplication is a bottleneck, use candidate blocking to reduce O(n²) comparisons before similarity scoring:```python-resolver = EntityResolver()-merged = resolver.resolve(entities, strategy="semantic_v2") # up to 7x faster+from semantica.deduplication import DuplicateDetector, EntityMerger++entities = [+ {"id": "1", "name": "Acme Corp", "type": "Company"},+ {"id": "2", "name": "Acme Corporation", "type": "Company"},+ {"id": "3", "name": "Globex", "type": "Company"},+]++# Fast candidate blocking for large entity sets+detector = DuplicateDetector(similarity_threshold=0.8)+duplicates = detector.detect_duplicates(entities, candidate_strategy="blocking_v2")++merger = EntityMerger()+merged = merger.merge_duplicates(entities, strategy="keep_most_complete")```-The `blocking_v2`, `hybrid_v2`, and `semantic_v2` strategies reduce O(n²) comparisons via candidate blocking before similarity scoring.+The `blocking_v2` and `hybrid_v2` candidate strategies filter candidate pairs before calculating fine-grained similarity.</Accordion>@@ -233,8 +255,8 @@ The `blocking_v2`, `hybrid_v2`, and `semantic_v2` strategies reduce O(n²) compa- **API keys**: store in environment variables or a secrets manager; never commit them to version control; rotate on a schedule- **Sensitive data**: use local embedding models (Ollama, HuggingFace) for PII or classified content; avoid sending sensitive data to external APIs without data handling agreements-- **Graph exports**: encrypt sensitive exports at rest; use the v0.5.0 SSRF-safe `base_url` validation when configuring custom LLM gateways-- **XML ingestion**: always use `XMLIngestor` (v0.5.0), which uses the XXE-safe lxml backend; never parse untrusted XML with the standard library parser+- **Graph exports**: encrypt sensitive exports at rest; use SSRF-safe `base_url` validation when configuring custom LLM gateways+- **XML ingestion**: always use `XMLIngestor`, which uses the XXE-safe lxml backend; never parse untrusted XML with the standard library parser- [Cookbook](/cookbook): interactive Jupyter notebooks from beginner to advanced.- [FAQ](/faq): common questions answered.