-
Notifications
You must be signed in to change notification settings - Fork 0
/
create_database.py
232 lines (193 loc) · 7.37 KB
/
create_database.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
# import logging
# from langchain_community.document_loaders import DirectoryLoader
# from langchain.text_splitter import RecursiveCharacterTextSplitter
# from langchain.schema import Document
# from langchain_openai import OpenAIEmbeddings
# from langchain_community.vectorstores import Chroma
# import openai
# from dotenv import load_dotenv
# import os
# import shutil
# from pymongo import MongoClient
# # Configure logging
# logging.basicConfig(level=logging.INFO)
# logger = logging.getLogger(__name__)
# # Load environment variables. Assumes that project contains .env file with API keys
# load_dotenv()
# # Set OpenAI API key
# openai.api_key = os.environ['OPENAI_API_KEY']
# CHROMA_PATH = "chroma"
# MONGO_URI = os.environ['MONGO_URI']
# DATABASE_NAME = "document_db"
# COLLECTION_NAME = "documents"
# client = MongoClient(MONGO_URI)
# db = client[DATABASE_NAME]
# collection = db[COLLECTION_NAME]
# def main():
# try:
# logger.info("Starting database update...")
# generate_data_store()
# logger.info("Database updated successfully.")
# except Exception as e:
# logger.error(f"Error updating database: {str(e)}")
# finally:
# logger.info("Exiting script.")
# def generate_data_store():
# documents = load_documents()
# chunks = split_text(documents)
# save_to_chroma(chunks)
# def load_documents():
# documents = []
# for doc in collection.find():
# documents.append(Document(page_content=doc["page_content"], metadata=doc["metadata"]))
# logger.info(f"Loaded {len(documents)} documents from MongoDB.")
# return documents
# def split_text(documents: list[Document]):
# text_splitter = RecursiveCharacterTextSplitter(
# chunk_size=500,
# chunk_overlap=250,
# length_function=len,
# add_start_index=True,
# )
# chunks = text_splitter.split_documents(documents)
# logger.info(f"Split {len(documents)} documents into {len(chunks)} chunks.")
# document = chunks[10]
# logger.debug(document.page_content)
# logger.debug(document.metadata)
# return chunks
# def save_to_chroma(chunks: list[Document]):
# # Clear out the database first.
# if os.path.exists(CHROMA_PATH):
# shutil.rmtree(CHROMA_PATH)
# logger.info(f"Cleared existing Chroma database at {CHROMA_PATH}.")
# # Create a new DB from the documents.
# db = Chroma.from_documents(
# chunks, OpenAIEmbeddings(), persist_directory=CHROMA_PATH
# )
# db.persist()
# logger.info(f"Saved {len(chunks)} chunks to {CHROMA_PATH}.")
# if __name__ == "__main__":
# main()
import logging
from colorlog import ColoredFormatter
from langchain_community.document_loaders import DirectoryLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
# from langchain.schema import Document
from langchain_core.documents import Document
from langchain_community.vectorstores import Chroma
from sentence_transformers import SentenceTransformer
import openai
from dotenv import load_dotenv
import os
import shutil
from pymongo import MongoClient
import tiktoken
from datetime import datetime
# Configure logging with colorlog
formatter = ColoredFormatter(
"%(log_color)s%(levelname)s:%(name)s:%(message)s", # Fixed format string
datefmt=None,
reset=True,
log_colors={
'DEBUG': 'cyan',
'INFO': 'green',
'WARNING': 'yellow',
'ERROR': 'red',
'CRITICAL': 'red,bg_white',
}
)
handler = logging.StreamHandler()
handler.setFormatter(formatter)
logger = logging.getLogger(__name__)
logger.addHandler(handler)
logger.setLevel(logging.INFO)
# Load environment variables. Assumes that project contains .env file with API keys
load_dotenv()
CHROMA_PATH = "chroma"
MONGO_URI = os.environ['MONGO_URI']
DATABASE_NAME = "document_db"
COLLECTION_NAME = "documents"
TOKEN_LOG_FILE = "token_log.txt"
client = MongoClient(MONGO_URI)
db = client[DATABASE_NAME]
collection = db[COLLECTION_NAME]
class SentenceTransformerEmbeddings:
def __init__(self, model_name: str):
self.model = SentenceTransformer(model_name)
def embed_documents(self, documents: list[Document]):
print("Embedding documents...")
return [self.model.encode(doc).tolist() for doc in documents]
def main():
try:
logger.info("Starting database update...")
generate_data_store()
logger.info("Database updated successfully.")
except Exception as e:
logger.error(f"Error updating database: {str(e)}")
finally:
logger.info("Exiting script.")
def generate_data_store():
documents = load_documents()
chunks = split_text(documents)
save_to_chroma(chunks)
def load_documents():
documents = []
for doc in collection.find():
# Log the raw data coming from the database
# logger.info(f"Raw document from MongoDB: {doc}")
# Ensure the document is correctly converted from JSON to Document object
if isinstance(doc, dict) and "page_content" in doc and "metadata" in doc:
documents.append(Document(page_content=doc["page_content"], metadata=doc["metadata"]))
else:
logger.warning(f"Skipping invalid document: {doc}")
# print(documents[0])
logger.info(f"Loaded {len(documents)} documents from MongoDB.")
return documents
def split_text(documents):
text_splitter = RecursiveCharacterTextSplitter(
chunk_size=500,
chunk_overlap=250,
length_function=len,
add_start_index=True,
)
chunks = text_splitter.split_documents(documents)
logger.info(f"Split {len(documents)} documents into {len(chunks)} chunks.")
logger.info(type(chunks[0]))
# Check the type of each chunk
for i, chunk in enumerate(chunks):
if isinstance(chunk, Document):
logger.debug(f"Chunk {i}: Document content: {chunk.page_content[:50]}")
else:
logger.error(f"Chunk {i} is not a Document object: {chunk}")
# Convert chunk back to Document if needed
chunks[i] = Document(page_content=str(chunk), metadata={})
# print(chunks[2])
return chunks
# def log_token_count(documents):
# encoding = tiktoken.encoding_for_model("gpt-3.5")
# total_tokens = sum(len(encoding.encode(doc.page_content)) for doc in documents)
# with open(TOKEN_LOG_FILE, "a") as log_file:
# log_file.write(f"{datetime.now()}: Total Tokens: {total_tokens}\n")
def save_to_chroma(chunks):
# Validate chunks before proceeding
for i, chunk in enumerate(chunks):
if not isinstance(chunk, Document):
logger.error(f"Invalid chunk at index {i}: {chunk}")
return
logger.debug(f"Chunk {i} - page_content: {chunk.page_content[:50]}, metadata: {chunk.metadata}")
if os.path.exists(CHROMA_PATH):
shutil.rmtree(CHROMA_PATH)
logger.info(f"Cleared existing Chroma database at {CHROMA_PATH}.")
embedding_model = SentenceTransformerEmbeddings('sentence-transformers/all-MiniLM-L6-v2')
try:
# print("Creating Chroma database...",chunks)
# print(chunks[0].page_content)
db = Chroma.from_documents(
chunks, embedding_model, persist_directory=CHROMA_PATH
)
db.persist()
logger.info(f"Saved {len(chunks)} chunks to {CHROMA_PATH}.")
except Exception as e:
logger.error(f"Failed to create Chroma database: {str(e)}")
if __name__ == "__main__":
main()