my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

my-project-public — gitlawb

#1	import hashlib
#2	import logging
#3	import os
#4	import ssl
#5	from typing import Any, Optional
#6
#7	import certifi
#8
#9	from embedchain.loaders.base_loader import BaseLoader
#10	from embedchain.utils.misc import clean_string
#11
#12	SLACK_API_BASE_URL = "https://www.slack.com/api/"
#13
#14	logger = logging.getLogger(__name__)
#15
#16
#17	class SlackLoader(BaseLoader):
#18	def __init__(self, config: Optional[dict[str, Any]] = None):
#19	super().__init__()
#20
#21	self.config = config if config else {}
#22
#23	if "base_url" not in self.config:
#24	self.config["base_url"] = SLACK_API_BASE_URL
#25
#26	self.client = None
#27	self._setup_loader(self.config)
#28
#29	def _setup_loader(self, config: dict[str, Any]):
#30	try:
#31	from slack_sdk import WebClient
#32	except ImportError as e:
#33	raise ImportError(
#34	"Slack loader requires extra dependencies. \
#35	Install with `pip install --upgrade embedchain[slack]`"
#36	) from e
#37
#38	if os.getenv("SLACK_USER_TOKEN") is None:
#39	raise ValueError(
#40	"SLACK_USER_TOKEN environment variables not provided. Check `https://docs.embedchain.ai/data-sources/slack` to learn more." # noqa:E501
#41	)
#42
#43	logger.info(f"Creating Slack Loader with config: {config}")
#44	# get slack client config params
#45	slack_bot_token = os.getenv("SLACK_USER_TOKEN")
#46	ssl_cert = ssl.create_default_context(cafile=certifi.where())
#47	base_url = config.get("base_url", SLACK_API_BASE_URL)
#48	headers = config.get("headers")
#49	# for Org-Wide App
#50	team_id = config.get("team_id")
#51
#52	self.client = WebClient(
#53	token=slack_bot_token,
#54	base_url=base_url,
#55	ssl=ssl_cert,
#56	headers=headers,
#57	team_id=team_id,
#58	)
#59	logger.info("Slack Loader setup successful!")
#60
#61	@staticmethod
#62	def _check_query(query):
#63	if not isinstance(query, str):
#64	raise ValueError(
#65	f"Invalid query passed to Slack loader, found: {query}. Check `https://docs.embedchain.ai/data-sources/slack` to learn more." # noqa:E501
#66	)
#67
#68	def load_data(self, query):
#69	self._check_query(query)
#70	try:
#71	data = []
#72	data_content = []
#73
#74	logger.info(f"Searching slack conversations for query: {query}")
#75	results = self.client.search_messages(
#76	query=query,
#77	sort="timestamp",
#78	sort_dir="desc",
#79	count=self.config.get("count", 100),
#80	)
#81
#82	messages = results.get("messages")
#83	num_message = len(messages)
#84	logger.info(f"Found {num_message} messages for query: {query}")
#85
#86	matches = messages.get("matches", [])
#87	for message in matches:
#88	url = message.get("permalink")
#89	text = message.get("text")
#90	content = clean_string(text)
#91
#92	message_meta_data_keys = ["iid", "team", "ts", "type", "user", "username"]
#93	metadata = {}
#94	for key in message.keys():
#95	if key in message_meta_data_keys:
#96	metadata[key] = message.get(key)
#97	metadata.update({"url": url})
#98
#99	data.append(
#100	{
#101	"content": content,
#102	"meta_data": metadata,
#103	}
#104	)
#105	data_content.append(content)
#106	doc_id = hashlib.md5((query + ", ".join(data_content)).encode()).hexdigest()
#107	return {
#108	"doc_id": doc_id,
#109	"data": data,
#110	}
#111	except Exception as e:
#112	logger.warning(f"Error in loading slack data: {e}")
#113	raise ValueError(
#114	f"Error in loading slack data: {e}. Check `https://docs.embedchain.ai/data-sources/slack` to learn more." # noqa:E501
#115	) from e
#116

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public