Spaces:
Running
Running
Update Classes/Owiki_Class.py
Browse files- Classes/Owiki_Class.py +91 -104
Classes/Owiki_Class.py
CHANGED
@@ -1,104 +1,91 @@
|
|
1 |
-
import os
|
2 |
-
import google.generativeai as genai
|
3 |
-
from langchain_google_genai import ChatGoogleGenerativeAI
|
4 |
-
from langchain_community.vectorstores import FAISS
|
5 |
-
from langchain_core.output_parsers import StrOutputParser
|
6 |
-
from langchain_core.runnables import RunnablePassthrough
|
7 |
-
from langchain.prompts import PromptTemplate
|
8 |
-
import json
|
9 |
-
import re
|
10 |
-
from Classes.Helper_Class import DB_Retriever
|
11 |
-
from typing import Optional
|
12 |
-
|
13 |
-
os.environ["GOOGLE_API_KEY"] = "AIzaSyBoghqvvnMMS4bA61LjQkkPNdIRetqk438"
|
14 |
-
genai.configure(api_key="AIzaSyBoghqvvnMMS4bA61LjQkkPNdIRetqk438")
|
15 |
-
|
16 |
-
class OWiki:
|
17 |
-
def __init__(self,**kwargs):
|
18 |
-
temperature = kwargs['temperature']
|
19 |
-
self.summary = kwargs['summary_length']
|
20 |
-
model = kwargs["model"]
|
21 |
-
self.db_loc = kwargs["db_loc"]
|
22 |
-
self.llm = ChatGoogleGenerativeAI(model=model,
|
23 |
-
temperature=temperature)
|
24 |
-
self.model_embedding = kwargs['model_embeddings']
|
25 |
-
|
26 |
-
|
27 |
-
def get_summary_template(self):
|
28 |
-
prompt = """Generate a summary for the following conversational data in less than {summary} lines.\nText:\n{text}\n\nSummary:"""
|
29 |
-
prompt_template = PromptTemplate(template = prompt,input_variables=['summary','text'])
|
30 |
-
return prompt_template
|
31 |
-
|
32 |
-
def create_sql_prompt_template(self,schemas):
|
33 |
-
prompt = """Write an SQL query for the following questions whose schemas are as follows.\nSQL Schema:"""
|
34 |
-
for table_name,table_schema in schemas.items():
|
35 |
-
prompt+= f"Table Name: {table_name}, Schema : {table_schema}\n\n"
|
36 |
-
prompt+= """\n\nQuestion:{question}\n\nAnswer:"""
|
37 |
-
prompt_template = PromptTemplate(template = prompt,input_variables=['question'])
|
38 |
-
return prompt_template
|
39 |
-
|
40 |
-
def create_prompt_for_OIC_bot(self):
|
41 |
-
template = """You are OIC(Oracle Integration Cloud) Bot.Follow chat instructions and answer the question based only on the following
|
42 |
-
Chat_instructions:
|
43 |
-
1. Response must contain Question Explaination along with Potential Solution Headings.
|
44 |
-
2. Response must contain all possible Error Scenarios if applicable along with a Summary Heading containing breif summary at the end.
|
45 |
-
|
46 |
-
Context:
|
47 |
-
{context}
|
48 |
-
|
49 |
-
Question: {question}
|
50 |
-
"""
|
51 |
-
prompt = PromptTemplate.from_template(template)
|
52 |
-
return prompt
|
53 |
-
|
54 |
-
def create_sql_agent(self,question,schemas):
|
55 |
-
prompt_template = self.create_sql_prompt_template(schemas)
|
56 |
-
chain = prompt_template | self.llm | StrOutputParser()
|
57 |
-
response = chain.invoke({"question":question})
|
58 |
-
response = self.format_llm_response(response)
|
59 |
-
return response
|
60 |
-
|
61 |
-
def generate_summary(self,text):
|
62 |
-
prompt_template = self.get_summary_template()
|
63 |
-
chain = prompt_template | self.llm | StrOutputParser()
|
64 |
-
response = chain.invoke({"text":text,"summary":self.summary})
|
65 |
-
return response
|
66 |
-
|
67 |
-
def format_llm_response(self,text):
|
68 |
-
bold_pattern = r"\*\*(.*?)\*\*"
|
69 |
-
italic_pattern = r"\*(.*?)\*"
|
70 |
-
code_pattern = r"```(.*?)```"
|
71 |
-
text = text.replace('\n', '<br>')
|
72 |
-
formatted_text = re.sub(code_pattern,"<pre><code>\\1</code></pre>",text)
|
73 |
-
formatted_text = re.sub(bold_pattern, "<b>\\1</b>", formatted_text)
|
74 |
-
formatted_text = re.sub(italic_pattern, "<i>\\1</i>", formatted_text)
|
75 |
-
return formatted_text
|
76 |
-
|
77 |
-
def search_from_db(self, query : str, chat_history : Optional[str] ) -> str :
|
78 |
-
db = DB_Retriever(self.db_loc,self.model_embedding)
|
79 |
-
retriever = db.retrieve(query)
|
80 |
-
prompt = self.create_prompt_for_OIC_bot()
|
81 |
-
chat_history = self.generate_summary(chat_history)
|
82 |
-
retrieval_chain = (
|
83 |
-
{"context": retriever, "question": RunnablePassthrough()}
|
84 |
-
| prompt
|
85 |
-
| self.llm
|
86 |
-
| StrOutputParser()
|
87 |
-
)
|
88 |
-
response = retrieval_chain.invoke(query)
|
89 |
-
|
90 |
-
return response
|
91 |
-
|
92 |
-
if __name__=="__main__":
|
93 |
-
with open("src/config.json",'r') as f:
|
94 |
-
hyperparameters = json.load(f)
|
95 |
-
a = OWiki(**hyperparameters)
|
96 |
-
# print(a.generate_summary("""User:What is ML?\nBot:Machine learning (ML) is a branch of
|
97 |
-
# and computer science that focuses on the using data and algorithms to enable AI to imitate the way that humans learn, gradually improving its accuracy.
|
98 |
-
|
99 |
-
# How does machine learning work?
|
100 |
-
# (link resides outside ibm.com) breaks out the learning system of a machine learning algorithm into three main parts.\nUser:How to integrate with Oracle\nUser:Explain what have you explained above\nBot:"""))
|
101 |
-
# print("*"*100)
|
102 |
-
# hyperparameters = {"User":" id INTEGER PRIMARY KEY AUTOINCREMENT, name TEXT NOT NULL, email TEXT UNIQUE","User1":" id1 INTEGER PRIMARY KEY AUTOINCREMENT, name TEXT NOT NULL, email TEXT UNIQUE"}
|
103 |
-
# print(a.create_sql_agent("Filter out common values in table 1 and 2 based on id",**hyperparameters))
|
104 |
-
print(a.search_from_db("What is Machine Learning","You can answer out of context as well"))
|
|
|
1 |
+
import os
|
2 |
+
import google.generativeai as genai
|
3 |
+
from langchain_google_genai import ChatGoogleGenerativeAI
|
4 |
+
from langchain_community.vectorstores import FAISS
|
5 |
+
from langchain_core.output_parsers import StrOutputParser
|
6 |
+
from langchain_core.runnables import RunnablePassthrough
|
7 |
+
from langchain.prompts import PromptTemplate
|
8 |
+
import json
|
9 |
+
import re
|
10 |
+
from Classes.Helper_Class import DB_Retriever
|
11 |
+
from typing import Optional
|
12 |
+
|
13 |
+
os.environ["GOOGLE_API_KEY"] = "AIzaSyBoghqvvnMMS4bA61LjQkkPNdIRetqk438"
|
14 |
+
genai.configure(api_key="AIzaSyBoghqvvnMMS4bA61LjQkkPNdIRetqk438")
|
15 |
+
|
16 |
+
class OWiki:
|
17 |
+
def __init__(self,**kwargs):
|
18 |
+
temperature = kwargs['temperature']
|
19 |
+
self.summary = kwargs['summary_length']
|
20 |
+
model = kwargs["model"]
|
21 |
+
self.db_loc = kwargs["db_loc"]
|
22 |
+
self.llm = ChatGoogleGenerativeAI(model=model,
|
23 |
+
temperature=temperature)
|
24 |
+
self.model_embedding = kwargs['model_embeddings']
|
25 |
+
|
26 |
+
|
27 |
+
def get_summary_template(self):
|
28 |
+
prompt = """Generate a summary for the following conversational data in less than {summary} lines.\nText:\n{text}\n\nSummary:"""
|
29 |
+
prompt_template = PromptTemplate(template = prompt,input_variables=['summary','text'])
|
30 |
+
return prompt_template
|
31 |
+
|
32 |
+
def create_sql_prompt_template(self,schemas):
|
33 |
+
prompt = """Write an SQL query for the following questions whose schemas are as follows.\nSQL Schema:"""
|
34 |
+
for table_name,table_schema in schemas.items():
|
35 |
+
prompt+= f"Table Name: {table_name}, Schema : {table_schema}\n\n"
|
36 |
+
prompt+= """\n\nQuestion:{question}\n\nAnswer:"""
|
37 |
+
prompt_template = PromptTemplate(template = prompt,input_variables=['question'])
|
38 |
+
return prompt_template
|
39 |
+
|
40 |
+
def create_prompt_for_OIC_bot(self):
|
41 |
+
template = """You are OIC(Oracle Integration Cloud) Bot.Follow chat instructions and answer the question based only on the following
|
42 |
+
Chat_instructions:
|
43 |
+
1. Response must contain Question Explaination along with Potential Solution Headings.
|
44 |
+
2. Response must contain all possible Error Scenarios if applicable along with a Summary Heading containing breif summary at the end.
|
45 |
+
|
46 |
+
Context:
|
47 |
+
{context}
|
48 |
+
|
49 |
+
Question: {question}
|
50 |
+
"""
|
51 |
+
prompt = PromptTemplate.from_template(template)
|
52 |
+
return prompt
|
53 |
+
|
54 |
+
def create_sql_agent(self,question,schemas):
|
55 |
+
prompt_template = self.create_sql_prompt_template(schemas)
|
56 |
+
chain = prompt_template | self.llm | StrOutputParser()
|
57 |
+
response = chain.invoke({"question":question})
|
58 |
+
response = self.format_llm_response(response)
|
59 |
+
return response
|
60 |
+
|
61 |
+
def generate_summary(self,text):
|
62 |
+
prompt_template = self.get_summary_template()
|
63 |
+
chain = prompt_template | self.llm | StrOutputParser()
|
64 |
+
response = chain.invoke({"text":text,"summary":self.summary})
|
65 |
+
return response
|
66 |
+
|
67 |
+
def format_llm_response(self,text):
|
68 |
+
bold_pattern = r"\*\*(.*?)\*\*"
|
69 |
+
italic_pattern = r"\*(.*?)\*"
|
70 |
+
code_pattern = r"```(.*?)```"
|
71 |
+
text = text.replace('\n', '<br>')
|
72 |
+
formatted_text = re.sub(code_pattern,"<pre><code>\\1</code></pre>",text)
|
73 |
+
formatted_text = re.sub(bold_pattern, "<b>\\1</b>", formatted_text)
|
74 |
+
formatted_text = re.sub(italic_pattern, "<i>\\1</i>", formatted_text)
|
75 |
+
return formatted_text
|
76 |
+
|
77 |
+
def search_from_db(self, query : str, chat_history : Optional[str] ) -> str :
|
78 |
+
db = DB_Retriever(self.db_loc,self.model_embedding)
|
79 |
+
retriever = db.retrieve(query)
|
80 |
+
prompt = self.create_prompt_for_OIC_bot()
|
81 |
+
chat_history = self.generate_summary(chat_history)
|
82 |
+
retrieval_chain = (
|
83 |
+
{"context": retriever, "question": RunnablePassthrough()}
|
84 |
+
| prompt
|
85 |
+
| self.llm
|
86 |
+
| StrOutputParser()
|
87 |
+
)
|
88 |
+
response = retrieval_chain.invoke(query)
|
89 |
+
response = self.format_llm_response(response)
|
90 |
+
return response
|
91 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|