[AUR-391, AUR-393] Add Document and DocumentReader base (#6)

* Declare BaseComponent

* Brainstorming base class for LLM call

* Define base LLM

* Add tests

* Clean telemetry environment for accurate testing

* Fix README

* Fix typing

* add base document reader

* update test

* update requirements

* Cosmetic change

* update requirements

* reformat

---------

Co-authored-by: trducng <trungduc1992@gmail.com>

knowledgehub/documents/base.py

@@ -0,0 +1,22 @@
+from haystack.schema import Document as HaystackDocument
+from llama_index.schema import Document as BaseDocument
+
+SAMPLE_TEXT = "A sample Document from kotaemon"
+
+
+class Document(BaseDocument):
+    """Base document class, mostly inherited from Document class from llama-index"""
+
+    @classmethod
+    def example(cls) -> "Document":
+        document = Document(
+            text=SAMPLE_TEXT,
+            metadata={"filename": "README.md", "category": "codebase"},
+        )
+        return document
+
+    def to_haystack_format(self) -> HaystackDocument:
+        """Convert struct to Haystack document format."""
+        metadata = self.metadata or {}
+        text = self.text
+        return HaystackDocument(content=text, meta=metadata)

knowledgehub/loaders/__init__.py

@@ -0,0 +1,3 @@
+from .base import AutoReader
+
+__all__ = ["AutoReader"]

knowledgehub/loaders/base.py

@@ -1,10 +1,26 @@
-class DocumentLoader:
-    """Document loader"""
+from pathlib import Path
+from typing import Any, List, Type, Union
+
+from llama_index import download_loader
+from llama_index.readers.base import BaseReader
+
+from ..documents.base import Document
 
 
-class TextManipulator:
-    """Text manipulation"""
+class AutoReader(BaseReader):
+    """General auto reader for a variety of files. (based on llama-hub)"""
 
+    def __init__(self, reader_type: Union[str, Type[BaseReader]]) -> None:
+        """Init reader using string identifier or class name from llama-hub"""
 
-class DocumentManipulator:
-    """Document manipulation"""
+        if isinstance(reader_type, str):
+            self._reader = download_loader(reader_type)()
+        else:
+            self._reader = reader_type()
+
+    def load_data(self, file: Union[Path, str], **kwargs: Any) -> List[Document]:
+        documents = self._reader.load_data(file=file, **kwargs)
+
+        # convert Document to new base class from kotaemon
+        converted_documents = [Document.from_dict(doc.to_dict()) for doc in documents]
+        return converted_documents