document_loaders — PDF, web, S3, Notion, ... unified Document model

PDF, web, S3, Notion, ... unified Document model

Module langchain_classic/document_loaders/ — PDF, web, S3, Notion, ... unified Document model. Part of the Retrieval & Data group.

🎯 Why This Module Exists

Bridges arbitrary data sources into a uniform Document(page_content, metadata) shape. 100+ loaders cover PDFs, HTML, Notion, Confluence, GitHub, Slack, S3, Google Drive, YouTube transcripts, etc. The BaseLoader.load() returns list[Document]; lazy_load() yields one at a time for streaming ingestion of huge corpora.

166

Source files

Sub-directories

Retrieval & Data

Architecture

flowchart LR
  subgraph Core["langchain_classic/document_loaders/"]
    BASE["Base abstraction"]
    IMPLS["Concrete implementations"]
    BASE --> IMPLS
  end
  subgraph Up["Upstream callers"]
    U1["chains/"]
    U2["agents/"]
  end
  subgraph Down["Downstream"]
    D1["langchain_core"]
    D2["External providers"]
  end
  Up --> Core
  Core --> Down

Data Flow

sequenceDiagram
  participant Caller as Caller
  participant Document as document_loaders
  participant Core as langchain_core
  participant Ext as External
  Caller->>Document: invoke(input)
  Note over Document: Validate & route
  Document->>Core: delegate primitive
  Core->>Ext: provider call
  Ext-->>Core: response
  Core-->>Document: typed output
  Document-->>Caller: result

Design Decisions

Decision	Choice Made	Alternatives	Why This Approach
Document shape	Dataclass: page_content (str) + metadata (dict)	Per-source type	Downstream splitters/embedders need a single uniform input
Lazy iteration	lazy_load() generator on top of load()	Always-eager	Some sources have GBs of content; eager would OOM
Optional dependencies	Per-loader import inside the file with helpful error	Hard requirements list	Hundreds of loaders × hundreds of clients would be unusable
Loader vs splitter separation	Loaders preserve source structure; splitting is a transformer	Loader splits inline	Splitter choice is application-specific (token-based, semantic, header-based)

Pairs with document_transformers/ (text splitters, redundancy filters) and feeds vectorstores/ ingestion.

Step-by-Step Walkthrough

Follow a typical operation through the document_loaders/ module:

Caller invokes the public surface — usually .invoke(input) on a Runnable, or a legacy __call__(inputs_dict) on a Chain subclass.
Pydantic validation — every entry point is a BaseModel; constructor args, input dicts, and tool args_schema are validated before any expensive work runs. Fail-fast on type errors.
CallbackManager.configure — a fresh manager is built per call, fans on_*_start events to all subscribers (LangSmith tracer, stdout handler, custom logging), and stamps a UUID4 run_id for tree reconstruction.
document_loaders-specific logic — the module's core method runs: prompt rendering, retrieval, tool execution, memory load, document split, etc. See the deep-dive sections above.
Provider call — typically a single HTTP request to OpenAI/Anthropic/etc., or a local SDK call. Retries are handled by tenacity with exponential backoff and a configurable max-attempts knob.
Streaming fan-out — if the provider streams tokens, each chunk fires on_llm_new_token through the callback manager so downstream observers (UI, log, tracer) see partial output.
OutputParser pass — if a parser is attached, the raw provider output is parsed into a structured Python object; OutputParserException may trigger a retry chain (RetryWithErrorOutputParser).
Memory write-back — for chains/agents that have a BaseMemory attached, save_context(inputs, outputs) persists the turn before returning to the caller.
on_*_end fan-out — the callback manager closes the run span; tracers serialise the Run tree and ship it to LangSmith asynchronously.
Result returned — caller receives a typed object (AIMessage, dict, BaseModel, list[Document], ...) matching the module's declared output schema.

What Could Go Wrong

Failure Mode	Trigger	Detection	Recovery
Provider 5xx / rate-limit	OpenAI/Anthropic/etc. transient error	HTTP status from SDK	tenacity retries with exponential backoff up to `max_retries`; then surface to caller
Output parser failure	LLM returned malformed JSON / wrong schema	`OutputParserException` in parser	`RetryWithErrorOutputParser` re-prompts with the parser error included; or surface
Pydantic validation	Caller passes wrong-shaped dict	`ValidationError` at `__init__` / `invoke`	Surface immediately — no fallback is correct here
Callback handler exception	User-supplied tracer raises	Try/except around each fan-out	Logged and swallowed by default; `raise_error=True` to opt in to propagation
Provider auth missing	`OPENAI_API_KEY` unset	SDK `AuthenticationError`	Surface with a helpful message; no retry (would just fail again)

Performance Profile

Hot Paths

Provider HTTP round-trip dominates wall-clock time (100ms–30s depending on model and prompt length).
Pydantic validation is repeated on every invoke; cached schema parsing keeps it under 1ms for typical inputs.
Callback fan-out is synchronous by default — a slow handler will slow every call. Async tracers (LangSmith) push work to a background thread to avoid blocking.

Resource Usage

Memory: token-history buffers (in memory/) and document caches (in retrieval pipelines) are the largest steady-state allocations.
I/O: bounded by provider rate limits; vector store I/O for retrieval-flavored modules.
Concurrency: async support via a* methods; agents and parallel chains use asyncio.gather for concurrent tool calls.

Directory Map

Source tree for document_loaders

document_loaders/
├── __init__.py
├── acreom.py
├── airbyte.py
├── airbyte_json.py
├── airtable.py
├── apify_dataset.py
├── arcgis_loader.py
├── arxiv.py
├── assemblyai.py
├── async_html.py
├── azlyrics.py
├── azure_ai_data.py
├── azure_blob_storage_container.py
├── azure_blob_storage_file.py
├── baiducloud_bos_directory.py
├── baiducloud_bos_file.py
├── base.py
├── base_o365.py
├── bibtex.py
├── bigquery.py
├── bilibili.py
├── blackboard.py
├── blob_loaders/
├── blockchain.py
├── brave_search.py
├── browserless.py
├── chatgpt.py
├── chromium.py
├── college_confidential.py
├── concurrent.py
├── confluence.py
├── conllu.py
├── couchbase.py
├── csv_loader.py
├── cube_semantic.py
├── datadog_logs.py
├── dataframe.py
├── diffbot.py
├── directory.py
├── discord.py
├── docugami.py
├── docusaurus.py
├── dropbox.py
├── duckdb_loader.py
├── email.py
├── epub.py
├── etherscan.py
├── evernote.py
├── excel.py
├── facebook_chat.py
├── fauna.py
├── figma.py
├── gcs_directory.py
├── gcs_file.py
├── generic.py
├── geodataframe.py
├── git.py
├── gitbook.py
├── github.py
├── google_speech_to_text.py
    ... (more files truncated)