API Reference
Last updated
Was this helpful?
Last updated
Was this helpful?
You can get a token from and try out our API.
Endpoint to list collections.
This endpoint retrieves a list of collections owned by the authenticated user.
Args: request: The request object containing authentication information.
Returns: A list of CollectionOut objects representing the collections owned by the authenticated user.
Raises: HTTPException: If there is an issue with the request or authentication.
OK
Retrieve a collection by its name.
Args: request: The request object containing authentication information. collection_name (str): The name of the collection to retrieve.
Returns: CollectionOut: The retrieved collection with its ID, name, and metadata.
Raises: HTTPException: If the collection is not found or the user is not authorized to access it.
Endpoint: GET /collections/{collection_name}
Tags: collections
Authentication: Bearer token required.
Delete a collection by its name.
This endpoint deletes a collection specified by the collection_name parameter.
The collection must belong to the authenticated user.
Args: request: The HTTP request object, which includes authentication information. collection_id (int): The ID of the collection to be deleted.
Returns: dict: A message indicating that the collection was deleted successfully.
Raises: HTTPException: If the collection does not exist or does not belong to the authenticated user.
No content
Retrieve a specific document from the user documents. Default collection is "default_collection". To get all documents, use collection_name="all".
Args: request: The HTTP request object. document_name (str): The ID of the document to retrieve. expand (Optional[str]): A comma-separated list of fields to expand in the response. If "pages" is included, the document's pages will be included.
Returns: DocumentOut: The retrieved document with its details.
Raises: HTTPException: If the document or collection is not found.
Example: GET /documents/{document_name}/?collection_name={collection_name}&expand=pages
default_collectionFetch a list of documents for a given collection.
This endpoint retrieves documents associated with a specified collection name. Optionally, it can expand the response to include pages of each document.
Args: request (Request): The request object. collection_name (Optional[str]): The name of the collection to fetch documents from. Defaults to "default_collection". Use "all" to fetch documents from all collections. expand (Optional[str]): A comma-separated string specifying additional fields to include in the response. If "pages" is included, the pages of each document will be included.
Returns: List[DocumentOut]: A list of documents with their details. If expanded, includes pages of each document.
Raises: HTTPException: If the collection or documents are not found.
Example: GET /documents/?collection_name=default_collection&expand=pages
default_collectionOK
Delete a document by its Name.
This endpoint deletes a document specified by the document_name parameter.
The document must belong to the authenticated user.
Args: request: The HTTP request object, which includes authentication information. collection_name (name): The name of the collection containing the document. Defaults to "default_collection". Use "all" to access all collections belonging to the user. document_name (int): The name of the document to be deleted.
Returns: dict: A message indicating that the document was deleted successfully.
Raises: HTTPException: If the document does not exist or does not belong to the authenticated user.
Example: DELETE /documents/delete-document/{document_name}/?collection_name={collection_name}
default_collectionNo content
Search for pages similar to a given query.
This endpoint allows the user to search for pages similar to a given query. The search is performed across all documents in the specified collection.
Args: request: The HTTP request object, which includes the user information. payload (QueryIn): The input data for the search, which includes the query string and collection ID.
Returns: QueryOut: The search results, including the query and a list of similar pages.
Raises: HttpError: If the collection does not exist or the query is invalid.
Example: POST /search/ { "query": "dog", "collection_name": "my_collection", "top_k": 3, "query_filter": { "on": "document", "key": "breed", "value": "collie", "lookup": "contains" } }
all3Search for pages similar to a given image.
This endpoint allows the user to search for pages similar to a given image. The search is performed across all documents in the specified collection.
Args: request: The HTTP request object, which includes the user information. payload (SearchImageIn): The input data for the search, which includes the image in base64 format and collection ID.
Returns: SearchImageOut: The search results, a list of similar pages.
Raises: HttpError: If the collection does not exist or the img_base64 is invalid.
Example: POST /search-image/ { "img_base64": "base64_string", "collection_name": "my_collection", "top_k": 3, "query_filter": { "on": "document", "key": "breed", "value": "collie", "lookup": "contains" } }
all3Create or update a document in a collection. Average latency is 7 seconds per page.
This endpoint allows the user to create or update a document in a collection. The document can be provided as a URL or a base64-encoded string. if the collection is not provided, a collection named "default_collection" will be used. if the collection is provided, it will be created if it does not exist.
Args: request: The HTTP request object, which includes the user information. payload (DocumentIn): The input data for creating or updating the document.
Returns: str: A message indicating that the document is being processed.
Raises: HttpError: If the document cannot be created or updated.
Example: POST /documents/upsert-document/ { "name": "my_document", "metadata": {"author": "John Doe"}, "collection": "my_collection", "url": "https://example.com/my_document.pdf, "wait": true # optional, if true, the response will be sent after waiting for the document to be processed. Otherwise, it will be done asynchronously. }
The name of the collection to which the document belongs. If not provided, the document will be added to the default_collection. Use 'all' to access all collections belonging to the user.
default_collectionfalsefalsePartially update a collection.
This endpoint allows for partial updates to a collection's details. Only the fields provided in the payload will be updated.
Args: request: The request object containing authentication details. collection_name (str): The name of the collection to be updated. payload (PatchCollectionIn): The payload containing the fields to be updated.
Returns: dict: A message indicating the collection was updated successfully.
Raises: HTTPException: If the collection is not found or the user is not authorized to update it.
Partially update a document.
This endpoint allows for partial updates to a document's details. Only the fields provided in the payload will be updated. If the URL is changed or base64 is provided, the document will be re-embedded. Otherwise, only the metadata and name will be updated.
Args: request: The request object containing authentication details. document_name (str): The name of the document to be updated. payload (DocumentInPatch): The payload containing the fields to be updated.
Returns: Tuple[int, DocumentOut] | Tuple[int, GenericError]: A tuple containing the status code and the updated document or an error message.
Raises: HTTPException: If the document is not found or the user is not authorized to update it.
The name of the collection to which the document belongs. If not provided, the document will be added to the default_collection. Use 'all' to access all collections belonging to the user.
default_collectionfalseUpload one file, converts to images and return their base64 encoded strings with 1-indexed page numberss.
Args: request: The HTTP request object. file UploadedFile): One uploaded file
Returns: List[FileOut]: A list of FileOut objects containing the base64 encoded strings of the images.
OK
Embed a list of documents.
This endpoint allows the user to embed a list of documents.
Args: request: The HTTP request object, which includes the user information. payload (EmbeddingsIn): The input data for embedding the documents.
Returns: EmbeddingsOut: The embeddings of the documents and metadata.
Raises: HttpError: If the documents cannot be embedded.