mirror of
https://github.com/hwchase17/langchain
synced 2024-11-08 07:10:35 +00:00
f59e5d48ed
My other [pull-request](https://github.com/langchain-ai/langchain/pull/5135) is too big to be acceptable. I propose another 'lite' version. I update only notebook to propose an integration with the external project [`langchain-googledrive`](https://github.com/pprados/langchain-googledrive). --------- Co-authored-by: Harrison Chase <hw.chase.17@gmail.com>
534 lines
17 KiB
Plaintext
534 lines
17 KiB
Plaintext
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "b0ed136e-6983-4893-ae1b-b75753af05f8",
|
|
"metadata": {},
|
|
"source": [
|
|
"# Google Drive\n",
|
|
"\n",
|
|
">[Google Drive](https://en.wikipedia.org/wiki/Google_Drive) is a file storage and synchronization service developed by Google.\n",
|
|
"\n",
|
|
"This notebook covers how to load documents from `Google Drive`. Currently, only `Google Docs` are supported.\n",
|
|
"\n",
|
|
"## Prerequisites\n",
|
|
"\n",
|
|
"1. Create a Google Cloud project or use an existing project\n",
|
|
"1. Enable the [Google Drive API](https://console.cloud.google.com/flows/enableapi?apiid=drive.googleapis.com)\n",
|
|
"1. [Authorize credentials for desktop app](https://developers.google.com/drive/api/quickstart/python#authorize_credentials_for_a_desktop_application)\n",
|
|
"1. `pip install --upgrade google-api-python-client google-auth-httplib2 google-auth-oauthlib`\n",
|
|
"\n",
|
|
"## 🧑 Instructions for ingesting your Google Docs data\n",
|
|
"By default, the `GoogleDriveLoader` expects the `credentials.json` file to be `~/.credentials/credentials.json`, but this is configurable using the `credentials_path` keyword argument. Same thing with `token.json` - `token_path`. Note that `token.json` will be created automatically the first time you use the loader.\n",
|
|
"\n",
|
|
"`GoogleDriveLoader` can load from a list of Google Docs document ids or a folder id. You can obtain your folder and document id from the URL:\n",
|
|
"* Folder: https://drive.google.com/drive/u/0/folders/1yucgL9WGgWZdM1TOuKkeghlPizuzMYb5 -> folder id is `\"1yucgL9WGgWZdM1TOuKkeghlPizuzMYb5\"`\n",
|
|
"* Document: https://docs.google.com/document/d/1bfaMQ18_i56204VaQDVeAFpqEijJTgvurupdEDiaUQw/edit -> document id is `\"1bfaMQ18_i56204VaQDVeAFpqEijJTgvurupdEDiaUQw\"`"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "6e40071c-3a65-4e26-b497-3e2be0bd86b9",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"!pip install --upgrade google-api-python-client google-auth-httplib2 google-auth-oauthlib"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "878928a6-a5ae-4f74-b351-64e3b01733fe",
|
|
"metadata": {
|
|
"tags": []
|
|
},
|
|
"outputs": [],
|
|
"source": [
|
|
"from langchain.document_loaders import GoogleDriveLoader"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "2216c83f-68e4-4d2f-8ea2-5878fb18bbe7",
|
|
"metadata": {
|
|
"tags": []
|
|
},
|
|
"outputs": [],
|
|
"source": [
|
|
"loader = GoogleDriveLoader(\n",
|
|
" folder_id=\"1yucgL9WGgWZdM1TOuKkeghlPizuzMYb5\",\n",
|
|
" # Optional: configure whether to recursively fetch files from subfolders. Defaults to False.\n",
|
|
" recursive=False,\n",
|
|
")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "8f3b6aa0-b45d-4e37-8c50-5bebe70fdb9d",
|
|
"metadata": {
|
|
"tags": []
|
|
},
|
|
"outputs": [],
|
|
"source": [
|
|
"docs = loader.load()"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "2721ba8a",
|
|
"metadata": {},
|
|
"source": [
|
|
"When you pass a `folder_id` by default all files of type document, sheet and pdf are loaded. You can modify this behaviour by passing a `file_types` argument "
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "2ff83b4c",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"loader = GoogleDriveLoader(\n",
|
|
" folder_id=\"1yucgL9WGgWZdM1TOuKkeghlPizuzMYb5\",\n",
|
|
" file_types=[\"document\", \"sheet\"],\n",
|
|
" recursive=False\n",
|
|
")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "d6b80931",
|
|
"metadata": {},
|
|
"source": [
|
|
"## Passing in Optional File Loaders\n",
|
|
"\n",
|
|
"When processing files other than Google Docs and Google Sheets, it can be helpful to pass an optional file loader to `GoogleDriveLoader`. If you pass in a file loader, that file loader will be used on documents that do not have a Google Docs or Google Sheets MIME type. Here is an example of how to load an Excel document from Google Drive using a file loader. "
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "94207e39",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"from langchain.document_loaders import GoogleDriveLoader\n",
|
|
"from langchain.document_loaders import UnstructuredFileIOLoader"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "a15fbee0",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"file_id = \"1x9WBtFPWMEAdjcJzPScRsjpjQvpSo_kz\"\n",
|
|
"loader = GoogleDriveLoader(\n",
|
|
" file_ids=[file_id],\n",
|
|
" file_loader_cls=UnstructuredFileIOLoader,\n",
|
|
" file_loader_kwargs={\"mode\": \"elements\"},\n",
|
|
")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "98410bda",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"docs = loader.load()"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "e3e72221",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"docs[0]"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "238cd06f",
|
|
"metadata": {},
|
|
"source": [
|
|
"You can also process a folder with a mix of files and Google Docs/Sheets using the following pattern:"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "0e2d093f",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"folder_id = \"1asMOHY1BqBS84JcRbOag5LOJac74gpmD\"\n",
|
|
"loader = GoogleDriveLoader(\n",
|
|
" folder_id=folder_id,\n",
|
|
" file_loader_cls=UnstructuredFileIOLoader,\n",
|
|
" file_loader_kwargs={\"mode\": \"elements\"},\n",
|
|
")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "b35ddcc6",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"docs = loader.load()"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "3cc141e0",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"docs[0]"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "e312268a",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": []
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "83ac576b-48c9-4aad-a35e-e978ea32f746",
|
|
"metadata": {},
|
|
"source": [
|
|
"## Extended usage\n",
|
|
"An external component can manage the complexity of Google Drive : `langchain-googledrive`\n",
|
|
"It's compatible with the ̀`langchain.document_loaders.GoogleDriveLoader` and can be used\n",
|
|
"in its place.\n",
|
|
"\n",
|
|
"To be compatible with containers, the authentication uses an environment variable ̀GOOGLE_ACCOUNT_FILE` to credential file (for user or service)."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "b94f7119-bc1e-4ca3-907f-9d81e837ac59",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"!pip install langchain-googledrive"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "c4c7474e-49cb-48a1-b3a0-77fba8e2dd70",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"folder_id='root'\n",
|
|
"#folder_id='1yucgL9WGgWZdM1TOuKkeghlPizuzMYb5'"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "8357f7f1-e2b1-41ef-8e38-48fcc3897dba",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"# Use the advanced version.\n",
|
|
"from langchain_googledrive.document_loaders import GoogleDriveLoader"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "16ab9d3d-1782-4cb9-ab56-d87edbb25a18",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"loader = GoogleDriveLoader(\n",
|
|
" folder_id=folder_id,\n",
|
|
" recursive=False,\n",
|
|
" num_results=2, # Maximum number of file to load\n",
|
|
")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "ebac43aa-dd64-4964-802a-a90172415fd1",
|
|
"metadata": {},
|
|
"source": [
|
|
"By default, all files with these mime-type can be converted to `Document`.\n",
|
|
"- text/text\n",
|
|
"- text/plain\n",
|
|
"- text/html\n",
|
|
"- text/csv\n",
|
|
"- text/markdown\n",
|
|
"- image/png\n",
|
|
"- image/jpeg\n",
|
|
"- application/epub+zip\n",
|
|
"- application/pdf\n",
|
|
"- application/rtf\n",
|
|
"- application/vnd.google-apps.document (GDoc)\n",
|
|
"- application/vnd.google-apps.presentation (GSlide)\n",
|
|
"- application/vnd.google-apps.spreadsheet (GSheet)\n",
|
|
"- application/vnd.google.colaboratory (Notebook colab)\n",
|
|
"- application/vnd.openxmlformats-officedocument.presentationml.presentation (PPTX)\n",
|
|
"- application/vnd.openxmlformats-officedocument.wordprocessingml.document (DOCX)\n",
|
|
"\n",
|
|
"It's possible to update or customize this. See the documentation of `GDriveLoader`.\n",
|
|
"\n",
|
|
"But, the corresponding packages must be installed."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "b4560f35-a37d-44e2-be0b-adaa245b3b3d",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"!pip install unstructured"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "6cb08da3-27df-46de-b60e-583bb7e31af4",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"for doc in loader.load():\n",
|
|
" print(\"---\")\n",
|
|
" print(doc.page_content.strip()[:60]+\"...\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "cd13d7d1-db7a-498d-ac98-76ccd9ad9019",
|
|
"metadata": {},
|
|
"source": [
|
|
"### Customize the search pattern\n",
|
|
"\n",
|
|
"All parameter compatible with Google [`list()`](https://developers.google.com/drive/api/v3/reference/files/list)\n",
|
|
"API can be set.\n",
|
|
"\n",
|
|
"To specify the new pattern of the Google request, you can use a `PromptTemplate()`.\n",
|
|
"The variables for the prompt can be set with `kwargs` in the constructor.\n",
|
|
"Some pre-formated request are proposed (use `{query}`, `{folder_id}` and/or `{mime_type}`):\n",
|
|
"\n",
|
|
"You can customize the criteria to select the files. A set of predefined filter are proposed:\n",
|
|
"| template | description |\n",
|
|
"| -------------------------------------- | --------------------------------------------------------------------- |\n",
|
|
"| gdrive-all-in-folder | Return all compatible files from a `folder_id` |\n",
|
|
"| gdrive-query | Search `query` in all drives |\n",
|
|
"| gdrive-by-name | Search file with name `query` |\n",
|
|
"| gdrive-query-in-folder | Search `query` in `folder_id` (and sub-folders if `recursive=true`) |\n",
|
|
"| gdrive-mime-type | Search a specific `mime_type` |\n",
|
|
"| gdrive-mime-type-in-folder | Search a specific `mime_type` in `folder_id` |\n",
|
|
"| gdrive-query-with-mime-type | Search `query` with a specific `mime_type` |\n",
|
|
"| gdrive-query-with-mime-type-and-folder | Search `query` with a specific `mime_type` and in `folder_id` |\n"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "81348d59-8fd6-45d4-9de3-5df5cff5c7e2",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"loader = GoogleDriveLoader(\n",
|
|
" folder_id=folder_id,\n",
|
|
" recursive=False,\n",
|
|
" template=\"gdrive-query\", # Default template to use\n",
|
|
" query=\"machine learning\",\n",
|
|
" num_results=2, # Maximum number of file to load\n",
|
|
" supportsAllDrives=False, # GDrive `list()` parameter\n",
|
|
")\n",
|
|
"for doc in loader.load():\n",
|
|
" print(\"---\")\n",
|
|
" print(doc.page_content.strip()[:60]+\"...\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "46c6ba5b-d4b1-4f0f-9801-5c1314021605",
|
|
"metadata": {},
|
|
"source": [
|
|
"You can customize your pattern."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "5a5a323b-8d96-46b7-b46a-fd69bd2c8e04",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"from langchain.prompts.prompt import PromptTemplate\n",
|
|
"loader = GoogleDriveLoader(\n",
|
|
" folder_id=folder_id,\n",
|
|
" recursive=False,\n",
|
|
" template=PromptTemplate(\n",
|
|
" input_variables=[\"query\", \"query_name\"],\n",
|
|
" template=\"fullText contains '{query}' and name contains '{query_name}' and trashed=false\",\n",
|
|
" ), # Default template to use\n",
|
|
" query=\"machine learning\",\n",
|
|
" query_name=\"ML\", \n",
|
|
" num_results=2, # Maximum number of file to load\n",
|
|
")\n",
|
|
"for doc in loader.load():\n",
|
|
" print(\"---\")\n",
|
|
" print(doc.page_content.strip()[:60]+\"...\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "375bb465-8f69-407b-94bd-ffa3718ef500",
|
|
"metadata": {},
|
|
"source": [
|
|
"#### Modes for GSlide and GSheet\n",
|
|
"The parameter mode accepts different values:\n",
|
|
"\n",
|
|
"- \"document\": return the body of each document\n",
|
|
"- \"snippets\": return the description of each file (set in metadata of Google Drive files).\n",
|
|
"\n",
|
|
"\n",
|
|
"The conversion can manage in Markdown format:\n",
|
|
"- bullet\n",
|
|
"- link\n",
|
|
"- table\n",
|
|
"- titles\n",
|
|
"\n",
|
|
"The parameter `gslide_mode` accepts different values:\n",
|
|
"\n",
|
|
"- \"single\" : one document with <PAGE BREAK>\n",
|
|
"- \"slide\" : one document by slide\n",
|
|
"- \"elements\" : one document for each elements.\n"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "7493d7b0-0600-49af-8107-7f4597c92de7",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"loader = GoogleDriveLoader(\n",
|
|
" template=\"gdrive-mime-type\",\n",
|
|
" mime_type=\"application/vnd.google-apps.presentation\", # Only GSlide files\n",
|
|
" gslide_mode=\"slide\",\n",
|
|
" num_results=2, # Maximum number of file to load\n",
|
|
")\n",
|
|
"for doc in loader.load():\n",
|
|
" print(\"---\")\n",
|
|
" print(doc.page_content.strip()[:60]+\"...\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "9bf338fb-02d7-452f-8679-c50419b13464",
|
|
"metadata": {},
|
|
"source": [
|
|
"The parameter `gsheet_mode` accepts different values:\n",
|
|
"- `\"single\"`: Generate one document by line\n",
|
|
"- `\"elements\"` : one document with markdown array and <PAGE BREAK> tags."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "469f5af0-67db-4f15-8aee-88cde480729b",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"loader = GoogleDriveLoader(\n",
|
|
" template=\"gdrive-mime-type\",\n",
|
|
" mime_type=\"application/vnd.google-apps.spreadsheet\", # Only GSheet files\n",
|
|
" gsheet_mode=\"elements\",\n",
|
|
" num_results=2, # Maximum number of file to load\n",
|
|
")\n",
|
|
"for doc in loader.load():\n",
|
|
" print(\"---\")\n",
|
|
" print(doc.page_content.strip()[:60]+\"...\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"id": "09acb864-e919-4add-9e06-deba6f7f0cd8",
|
|
"metadata": {},
|
|
"source": [
|
|
"### Advanced usage\n",
|
|
"All Google File have a 'description' in the metadata. This field can be used to memorize a summary of the document or others indexed tags (See method `lazy_update_description_with_summary()`).\n",
|
|
"\n",
|
|
"If you use the `mode=\"snippet\"`, only the description will be used for the body. Else, the `metadata['summary']` has the field.\n",
|
|
"\n",
|
|
"Sometime, a specific filter can be used to extract some information from the filename, to select some files with specific criteria. You can use a filter.\n",
|
|
"\n",
|
|
"Sometimes, many documents are returned. It's not necessary to have all documents in memory at the same time. You can use the lazy versions of methods, to get one document at a time. It's better to use a complex query in place of a recursive search. For each folder, a query must be applied if you activate `recursive=True`."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "a5e9c8eb-a266-4ae6-a760-d7826a0aa7c5",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"import os\n",
|
|
"loader = GoogleDriveLoader(\n",
|
|
" gdrive_api_file=os.environ[\"GOOGLE_ACCOUNT_FILE\"],\n",
|
|
" num_results=2,\n",
|
|
" template=\"gdrive-query\",\n",
|
|
" filter=lambda search, file: \"#test\" not in file.get('description',''),\n",
|
|
" query='machine learning',\n",
|
|
" supportsAllDrives=False,\n",
|
|
" )\n",
|
|
"for doc in loader.load():\n",
|
|
" print(\"---\")\n",
|
|
" print(doc.page_content.strip()[:60]+\"...\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"id": "51efa73a-4e2d-4f9c-abaf-6c9bde2ff69d",
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": []
|
|
}
|
|
],
|
|
"metadata": {
|
|
"kernelspec": {
|
|
"display_name": "Python 3 (ipykernel)",
|
|
"language": "python",
|
|
"name": "python3"
|
|
},
|
|
"language_info": {
|
|
"codemirror_mode": {
|
|
"name": "ipython",
|
|
"version": 3
|
|
},
|
|
"file_extension": ".py",
|
|
"mimetype": "text/x-python",
|
|
"name": "python",
|
|
"nbconvert_exporter": "python",
|
|
"pygments_lexer": "ipython3",
|
|
"version": "3.10.12"
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 5
|
|
}
|