API Reference — Overview

On this page

This section documents every public class in document_loaders, organized one page per module/file. Each page covers every public class defined in that module, including constructor signature, attributes, public methods, internal workflow, exceptions, thread safety, and examples drawn directly from the implementation.

Public API surface

This table mirrors document_loaders/__init__'s __all__ plus the symbols added by the later try/except and direct-import blocks at the bottom of that file.

SymbolKindModuleNotes
LoadedDocumentdataclassbase_loaderThe universal document model
BaseDocumentLoaderabstract classbase_loaderRoot abstract base for all loaders
BaseFileLoaderabstract classbase_loaderAbstract base for file-path-based loaders
LoaderConfigdataclassconfig_loaderMaster config aggregate (unused by loader constructors)
LoaderTypeenumconfig_loaderFormat identifiers
EXTENSION_MAPdictconfig_loaderExtension → LoaderType dispatch table
TextLoaderConfigdataclassconfig_loader—
PDFLoaderConfigdataclassconfig_loader—
DocxLoaderConfigdataclassconfig_loader—
CSVLoaderConfigdataclassconfig_loader—
JSONLoaderConfigdataclassconfig_loader—
HTMLLoaderConfigdataclassconfig_loader—
WebLoaderConfigdataclassconfig_loader—
DirectoryLoaderConfigdataclassconfig_loader—
TextLoaderclasstext_loader.txt files
MarkdownLoaderclasstext_loader.md files
PDFLoaderclasspdf_loader.pdf files
DocxLoaderclassdocx_loader.docx/.doc files
CSVLoaderclasscsv_loader.csv/.tsv files
ExcelLoaderclasscsv_loader.xlsx/.xls/.xlsm files
JSONLoaderclassjson_loader.json files
JSONLinesLoaderclassjson_loader.jsonl/.ndjson files
HTMLLoaderclasshtml_loader.html/.htm files
HTMLStringLoaderclasshtml_loaderIn-memory HTML strings
WebLoaderclassweb_loaderSingle URL
MultiURLLoaderclassweb_loaderConcurrent multi-URL
DirectoryLoaderclassdirectory_loaderRecursive directory ingestion
AutoLoaderclassauto_loaderFormat auto-detection dispatcher
ImageOCRLoader†classmultimodal_loaderImage OCR
PDFMultimodalLoader†classmultimodal_loaderPDF text + embedded-image OCR
TableExtractorLoader†classmultimodal_loaderHTML table → Markdown
GraphLoaderclassgraph_loaderKnowledge graph files
GraphLoaderConfigdataclassgraph_loader—
InMemoryGraphLoaderclassgraph_loaderIn-memory graph data
TableLoaderclasstable_loaderFormat-agnostic table extraction
TableLoaderConfigdataclasstable_loader—
InMemoryTableLoaderclasstable_loaderIn-memory table data

† multimodal_loader symbols are imported inside a try/except ImportError block in __init__.py. If the module fails to import (e.g. because zerda.document_loaders isn't resolvable as an absolute import path in your installation), these names are silently not added to the package namespace rather than raising at import time.

Pages

PageClasses covered
LoadedDocument & base classesLoadedDocument, BaseDocumentLoader, BaseFileLoader
Text & MarkdownTextLoader, MarkdownLoader
PDFPDFLoader
WordDocxLoader
CSV & ExcelCSVLoader, ExcelLoader
JSON & JSONLJSONLoader, JSONLinesLoader
HTMLHTMLLoader, HTMLStringLoader
WebWebLoader, MultiURLLoader
DirectoryDirectoryLoader
AutoAutoLoader
GraphGraphLoader, GraphLoaderConfig, InMemoryGraphLoader
TableTableLoader, TableLoaderConfig, InMemoryTableLoader

Conventions used across these pages

  • Thread safety statements describe the implementation as analyzed; they are not load-tested guarantees.
  • Version History sections state "cannot be determined from the repository" wherever no changelog, __version__, or git history was available — which is the case for the entire package as provided.
  • Inline # FIX: comments found in the source (e.g. in base_loader, csv_loader, docx_loader, directory_loader) describe bugs that were fixed in past revisions of this code; where relevant, this documentation calls these out as Implementation Notes since they explain non-obvious design choices.

Related pages