File size: 4,105 Bytes
4c95dc7
 
 
 
 
 
 
31f9732
 
4c95dc7
d523035
31f9732
d7ef377
 
4c95dc7
 
 
 
 
 
 
d7ef377
 
4c95dc7
 
 
8187b01
 
 
4c95dc7
b2f993e
 
 
 
 
 
 
 
 
 
 
 
 
 
4c95dc7
 
 
 
 
 
 
 
d523035
4c95dc7
 
 
 
 
 
 
 
 
d523035
4c95dc7
 
 
 
 
 
 
 
 
d523035
4c95dc7
 
8187b01
 
 
 
b2f993e
31f9732
 
4c95dc7
b2f993e
4c95dc7
8187b01
 
 
 
b2f993e
 
 
 
d523035
31f9732
 
 
 
 
 
b2f993e
 
 
 
 
 
31f9732
8187b01
 
 
 
 
 
b2f993e
 
 
 
 
 
 
 
 
 
 
0f09cc9
8187b01
31f9732
 
0f09cc9
31f9732
 
8187b01
 
 
 
31f9732
 
 
 
 
4c95dc7
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
from langchain_anthropic import ChatAnthropic
from langchain_openai import ChatOpenAI
from langchain.callbacks.manager import CallbackManager
from langchain.callbacks.tracers import LangChainTracer
from langchain_huggingface.embeddings import HuggingFaceEmbeddings
from langchain_experimental.text_splitter import SemanticChunker
from langchain_openai.embeddings import OpenAIEmbeddings
from langchain_qdrant import QdrantVectorStore, Qdrant
from langchain.retrievers.contextual_compression import ContextualCompressionRetriever
from qdrant_client import QdrantClient
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langchain_cohere import CohereRerank
from langchain_core.globals import set_llm_cache
from langchain_core.caches import InMemoryCache
import constants
import os

os.environ["LANGCHAIN_API_KEY"] = constants.LANGCHAIN_API_KEY
os.environ["LANGCHAIN_TRACING_V2"] = str(constants.LANGCHAIN_TRACING_V2)
os.environ["LANGCHAIN_ENDPOINT"] = constants.LANGCHAIN_ENDPOINT

set_llm_cache(InMemoryCache())

tracer = LangChainTracer()
callback_manager = CallbackManager([tracer])

########################
### Chat Models      ###
########################

#opus3 = ChatAnthropic(
#    api_key=constants.ANTRHOPIC_API_KEY, 
#    temperature=0, 
#    model='claude-3-opus-20240229',
#    callbacks=callback_manager
#)
#
#sonnet35 = ChatAnthropic(
#    api_key=constants.ANTRHOPIC_API_KEY, 
#    temperature=0, 
#    model='claude-3-5-sonnet-20240620',
#    max_tokens=4096,
#    callbacks=callback_manager
#)

gpt4 = ChatOpenAI(
    model="gpt-4",
    temperature=0,
    max_tokens=None,
    timeout=None,
    max_retries=2,
    api_key=constants.OPENAI_API_KEY,
    callbacks=callback_manager
)

gpt4o = ChatOpenAI(
    model="gpt-4o",
    temperature=0,
    max_tokens=None,
    timeout=None,
    max_retries=2,
    api_key=constants.OPENAI_API_KEY,
    callbacks=callback_manager
)

gpt4o_mini = ChatOpenAI(
    model="gpt-4o-mini",
    temperature=0,
    max_tokens=None,
    timeout=None,
    max_retries=2,
    api_key=constants.OPENAI_API_KEY,
    callbacks=callback_manager
)

########################
### Embedding Models ###
########################

#basic_embeddings = HuggingFaceEmbeddings(model_name="snowflake/snowflake-arctic-embed-l")

tuned_embeddings = HuggingFaceEmbeddings(model_name="CoExperiences/snowflake-l-marketing-tuned")

#te3_small = OpenAIEmbeddings(api_key=constants.OPENAI_API_KEY, model="text-embedding-3-small")

#######################
### Text Splitters  ###
#######################

#semanticChunker = SemanticChunker(
#    te3_small,
#    breakpoint_threshold_type="percentile"
#)

semanticChunker_tuned = SemanticChunker(
    tuned_embeddings,
    breakpoint_threshold_type="percentile",
    breakpoint_threshold_amount=85
)

#RCTS = RecursiveCharacterTextSplitter(
#    # Set a really small chunk size, just to show.
#    chunk_size=500,
#    chunk_overlap=25,
#    length_function=len,
#)

#######################
###  Vector Stores  ###
#######################

qdrant_client = QdrantClient(url=constants.QDRANT_ENDPOINT, api_key=constants.QDRANT_API_KEY)

#semantic_Qdrant_vs = QdrantVectorStore(
#    client=qdrant_client,
#    collection_name="docs_from_ripped_urls",
#    embedding=te3_small
#)
#
#rcts_Qdrant_vs = QdrantVectorStore(
#    client=qdrant_client,
#    collection_name="docs_from_ripped_urls_recursive",
#    embedding=te3_small
#)
collection_name="docs_from_ripped_urls_semantic_tuned"

semantic_tuned_Qdrant_vs = QdrantVectorStore(
    client=qdrant_client,
    collection_name=collection_name,
    embedding=tuned_embeddings
)

#######################
###  Retrievers     ###
#######################
semantic_tuned_retriever = semantic_tuned_Qdrant_vs.as_retriever(search_kwargs={"k" : 10})

compressor = CohereRerank(model="rerank-english-v3.0")
compression_retriever = ContextualCompressionRetriever(
    base_compressor=compressor, base_retriever=semantic_tuned_retriever
)