Hugging Face's logo

Spaces:
snehasquasher
/
spur-chatbot
Sleeping

App Files Community
spur-chatbot
File size: 8,258 Bytes 
b0d4092
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
233
234
235
236
237
238
239
240
241
242
243
 
import os
import openai
from langchain.text_splitter import CharacterTextSplitter
from langchain.document_loaders import UnstructuredFileLoader
from langchain.vectorstores.faiss import FAISS
from langchain.embeddings import OpenAIEmbeddings
from langchain.document_loaders import DirectoryLoader
from langchain.document_loaders import TextLoader
from langchain.document_loaders import CSVLoader
from langchain.document_loaders import PyPDFLoader
from langchain.document_loaders import UnstructuredWordDocumentLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter, Language
from langchain.vectorstores import Chroma
from langchain.document_loaders import NotionDBLoader 
from langchain.vectorstores.utils import filter_complex_metadata
import pickle
from Constants import *
from apiKey import *
from db_types import *
from utilities import transform_complex_metadata

def createChromaFromNotiondb(documents, embeddings) : 
    vectordb = Chroma(persist_directory=NOTION_PERSIST_DIRECTORY, embedding_function=embeddings,                                   
                      collection_name=NOTION_COLLECTION_NAME)
    print("Checking for existing collection count "+str(vectordb._collection.count()))
    if (vectordb._collection.count()== 0):
        print("Transforming notion collection "+ NOTION_COLLECTION_NAME)
        documents = transform_complex_metadata(documents)
        print("Creating notion database")
        vectordb = Chroma.from_documents(documents=documents, embedding=embeddings, persist_directory=NOTION_PERSIST_DIRECTORY, collection_name=NOTION_COLLECTION_NAME)
        vectordb.persist()
        print("Count of Notion collections: " + str(vectordb._collection.count()))
    else :
        print("Count of Notion collections: " + str(vectordb._collection.count()))
        
def createChromadb(documents, embeddings) : 
    vectordb = Chroma(persist_directory=CHROMA_PERSIST_DIRECTORY, embedding_function=embeddings,                                   
                      collection_name=CHROMA_COLLECTION_NAME)
    if (vectordb._collection.count()== 0):
        print("Creating chromadb")
        vectordb = Chroma.from_documents(documents=documents, embedding=embeddings, persist_directory=CHROMA_PERSIST_DIRECTORY, collection_name=CHROMA_COLLECTION_NAME)
        vectordb.persist()
        print("Count of collections: " + str(vectordb._collection.count()))
    else :
        print("Count of collections: " + str(vectordb._collection.count()))
    
def createFaissVectorstore(documents, embeddings) :
    print("Creating vectorstore...")
    vectorstore = FAISS.from_documents(documents, embeddings)
    with open("myvectorstore.pkl", "wb") as f:
        pickle.dump(vectorstore, f)
        
def enrichMetada(docs):
    
    for doc in docs: 
        for m in custom_meta_data:
            if (doc.metadata["source"] != ""):
                if ((m.get("name"))in doc.metadata["source"] ):
                     doc.metadata["name"] = m.get("name")
                     doc.metadata["profile"] = m.get("profile")
                     doc.metadata["creationYear"] = m.get("creationYear")
                     doc.metadata["topics"] = m.get("topics")

class MyLoader:
    def __init__(self, file_path, **kwargs):
        if file_path.endswith('.docx'):
          self.loader = UnstructuredWordDocumentLoader(file_path, **kwargs)
        elif file_path.endswith('.pdf'):
          self.loader = PyPDFLoader(file_path, **kwargs)
        elif file_path.endswith('.csv'):
            self.loader = CSVLoader(file_path, **kwargs)
        else:
            self.loader = TextLoader(file_path, **kwargs)
    
    def load(self):
        return self.loader.load()
    
custom_meta_data = [ 
    {
        "name":"Tanmay Chopra",
        "profile":"https://www.linkedin.com/in/tanmayc98/",
        "creationYear":"2023",
        "topics":"Pinecone",
    },
    {
        "name":"Neal Patel",
        "profile":"https://www.linkedin.com/in/nealpatel112/",
        "creationYear":"2023",
        "topics" :"Core - Model",
    },
    {
        "name":"Navid",
        "profile":"https://www.linkedin.com/in/Navid",
        "creationYear":"2022",
        "topics":"LLM",
    },
   {
        "name":"Josua Krause",
        "profile":"https://www.linkedin.com/in/Josua",
        "creationYear":"2022",
        "topics":"vector databases",
    },
    {
        "name":"Jay Zhong",
        "profile":"https://www.linkedin.com/in/Jay",
        "creationYear":"2021",
        "topics" : "LLM",
    },
    {
        "name":"Evan",
        "profile":"https://www.linkedin.com/in/Evan",
        "creationYear":"2021",
        "topics":"OpenAI",
    },
       {
        "name":"Siva_values",
        "profile":"https://www.linkedin.com/Siva",
        "creationYear":"2023",
        "topics":"Personal goals"
    },
    ]  

custom_meta_data = [ 
    {
        "name":"Tanmay Chopra",
        "profile":"https://www.linkedin.com/in/tanmayc98/",
        "creationYear":"2023",
        "topics":"Pinecone",
    },
    {
        "name":"Neal Patel",
        "profile":"https://www.linkedin.com/in/nealpatel112/",
        "creationYear":"2023",
        "topics" :"Core - Model",
    },
    {
        "name":"Navid",
        "profile":"https://www.linkedin.com/in/Navid",
        "creationYear":"2022",
        "topics":"LLM",
    },
   {
        "name":"Josua Krause",
        "profile":"https://www.linkedin.com/in/Josua",
        "creationYear":"2022",
        "topics":"vector databases",
    },
    {
        "name":"Jay Zhong",
        "profile":"https://www.linkedin.com/in/Jay",
        "creationYear":"2021",
        "topics" : "LLM",
    },
    {
        "name":"Evan",
        "profile":"https://www.linkedin.com/in/Evan",
        "creationYear":"2021",
        "topics":"OpenAI",
    },
       {
        "name":"Siva_values",
        "profile":"https://www.linkedin.com/Siva",
        "creationYear":"2023",
        "topics":"Personal goals"
    },
    ]  
custom_meta_data = [ 
    {
        "name":"Tanmay Chopra",
        "profile":"https://www.linkedin.com/in/tanmayc98/",
        "creationYear":"2023",
        "topics":"Pinecone",
    },
    {
        "name":"Neal Patel",
        "profile":"https://www.linkedin.com/in/nealpatel112/",
        "creationYear":"2023",
        "topics" :"Core - Model",
    },
    {
        "name":"Navid",
        "profile":"https://www.linkedin.com/in/Navid",
        "creationYear":"2022",
        "topics":"LLM",
    },
   {
        "name":"Josua Krause",
        "profile":"https://www.linkedin.com/in/Josua",
        "creationYear":"2022",
        "topics":"vector databases",
    },
    {
        "name":"Jay Zhong",
        "profile":"https://www.linkedin.com/in/Jay",
        "creationYear":"2021",
        "topics" : "LLM",
    },
    {
        "name":"Evan",
        "profile":"https://www.linkedin.com/in/Evan",
        "creationYear":"2021",
        "topics":"OpenAI",
    },
       {
        "name":"Siva_values",
        "profile":"https://www.linkedin.com/Siva",
        "creationYear":"2023",
        "topics":"Personal goals"
    },
    ]  

def ingestData():
    os.environ['OPENAI_API_KEY'] =OPENAI_API_KEY
    print("Loading data...")
    
    embeddings = OpenAIEmbeddings()

    if (DB_TYPE == DBTypes['FAISS'].value or DB_TYPE == DBTypes['CHROMA'].value) :
        loader = DirectoryLoader(DATA_DIRECTORY, glob="**/*.*", loader_cls=MyLoader)
        print("Loading directory")
        docs = loader.load()

        text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=0)

        enrichMetada(docs)
        print("splitting documents")
        documents = (text_splitter.split_documents(docs))
        if (DB_TYPE == DBTypes['FAISS']):
            createFaissVectorstore(documents, embeddings)
        elif (DB_TYPE == DBTypes['CHROMA'].value) : 
            createChromadb(documents, embeddings)
    elif (DB_TYPE == DBTypes['NOTION'].value):
        loader = NotionDBLoader(
            integration_token=NOTION_API_KEY,
            database_id=NOTION_DB,
            request_timeout_sec=30,  # optional, defaults to 10
        )

        documents = loader.load()
        createChromaFromNotiondb(documents, embeddings)
    
#ingestData()