feat: [DocAgent] Add Dynamic RAG

[priyansh4320]

Depends on #2097

Why are these changes needed?

Extending DocAgent Beyond Vector RAG to Graph RAG

The current DocAgent implementation in ag2 relies solely on vector-based retrieval-augmented generation (RAG), which while effective for semantic similarity searches, has inherent limitations in capturing complex relational information and multi-hop reasoning scenarios. Vector RAG excels at finding documents with similar embeddings but struggles with queries that require understanding relationships between entities, temporal sequences, or hierarchical structures within the knowledge base. By extending DocAgent to support Graph RAG, you could leverage knowledge graphs to represent documents as interconnected nodes and relationships, enabling more sophisticated reasoning patterns like "find all research papers that cite work by authors who collaborated with X" or "trace the evolution of a concept across multiple related documents."

Summary

• Updated DocAgent for Neo4j Graph Rag Capabilities.
• Updated Configuration for Dynamic RAG selection with a new parameter called rag_config.
• added result aggregator for when both RAGs are enabled.
• Sample Configuration for Dynamic RAG

llm_config = LLMConfig(
    model="gpt-5-nano",
    api_key=os.getenv("OPENAI_API_KEY"),
    api_type="openai",
)

doc_agent = DocAgent(
    llm_config=llm_config,
    rag_config={
        "graph": {
            "host": "bolt://localhost",
            "port": 7687,
            "username": "neo4j",
            "password": "admin1234",
        },
        "vector": {"engine": VectorChromaQueryEngine(collection_name="test7")},
    },
)

res = doc_agent.run(
    message="ingest the file test/agentchat/contrib/graph_rag/research_paper.pdf also ingest test/agentchat/contrib/graph_rag/Toast_financial_report.pdf? and who are theauthor of the research paper?",
    max_turns=1,
)
res.process()

Updates on the design

Related issue number

#2113

Checks

• I've included any doc changes needed for https://docs.ag2.ai/. See https://docs.ag2.ai/latest/docs/contributor-guide/documentation/ to build and test documentation locally.
• I've added tests (if relevant) corresponding to the changes introduced in this PR.
• I've made sure all auto checks have passed.

[marklysze]

@bassilkhilo-ag2 can you please try this out

[codecov]

Codecov Report

❌ Patch coverage is 40.92072% with 231 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
...agents/experimental/document_agent/task_manager.py	14.63%	140 Missing ⚠️
...ents/experimental/document_agent/document_agent.py	8.45%	65 Missing ⚠️
.../experimental/document_agent/task_manager_utils.py	87.24%	13 Missing and 6 partials ⚠️
...agents/experimental/document_agent/parser_utils.py	0.00%	7 Missing ⚠️

Files with missing lines	Coverage Δ
...agents/experimental/document_agent/parser_utils.py	23.61% <0.00%> (-71.85%)	⬇️
.../experimental/document_agent/task_manager_utils.py	87.24% <87.24%> (ø)
...ents/experimental/document_agent/document_agent.py	25.66% <8.45%> (+4.58%)	⬆️
...agents/experimental/document_agent/task_manager.py	14.63% <14.63%> (ø)

... and 40 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.