Update app.py
Browse files
app.py
CHANGED
@@ -1,6 +1,7 @@
|
|
1 |
import os
|
2 |
import logging
|
3 |
import asyncio
|
|
|
4 |
import gradio as gr
|
5 |
from huggingface_hub import InferenceClient
|
6 |
from langchain.embeddings import HuggingFaceEmbeddings
|
@@ -55,7 +56,16 @@ def create_web_search_vectors(search_results):
|
|
55 |
logging.info(f"Created vectors for {len(documents)} search results.")
|
56 |
return FAISS.from_documents(documents, embed)
|
57 |
|
58 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
59 |
search_results = duckduckgo_search(query)
|
60 |
|
61 |
if not search_results:
|
@@ -66,14 +76,7 @@ async def get_response_with_search(query, system_prompt, model, use_embeddings,
|
|
66 |
sources = [result['href'] for result in search_results if 'href' in result]
|
67 |
source_list_str = "\n".join(sources)
|
68 |
|
69 |
-
|
70 |
-
web_search_database = create_web_search_vectors(search_results)
|
71 |
-
retriever = web_search_database.as_retriever(search_kwargs={"k": 5})
|
72 |
-
relevant_docs = retriever.get_relevant_documents(query)
|
73 |
-
context = "\n".join([doc.page_content for doc in relevant_docs])
|
74 |
-
else:
|
75 |
-
context = "\n".join([f"{result['title']}\n{result['body']}" for result in search_results])
|
76 |
-
|
77 |
logging.info(f"Context created for query: {query}")
|
78 |
|
79 |
user_message = f"""Using the following context from web search results:
|
@@ -81,9 +84,6 @@ async def get_response_with_search(query, system_prompt, model, use_embeddings,
|
|
81 |
|
82 |
Write a detailed and complete research document that fulfills the following user request: '{query}'."""
|
83 |
|
84 |
-
client = InferenceClient(model, token=huggingface_token)
|
85 |
-
full_response = ""
|
86 |
-
|
87 |
messages = [
|
88 |
{"role": "system", "content": system_prompt},
|
89 |
{"role": "user", "content": user_message}
|
@@ -92,50 +92,38 @@ Write a detailed and complete research document that fulfills the following user
|
|
92 |
if history:
|
93 |
messages = history + messages
|
94 |
|
95 |
-
|
96 |
-
|
97 |
-
try:
|
98 |
-
response_stream = client.chat_completion(
|
99 |
-
messages=messages,
|
100 |
-
max_tokens=6000,
|
101 |
-
temperature=temperature,
|
102 |
-
stream=True,
|
103 |
-
top_p=0.8,
|
104 |
-
)
|
105 |
-
|
106 |
-
if response_stream is None:
|
107 |
-
logging.error(f"API call {call + 1} returned None")
|
108 |
-
yield "The API returned an empty response. Please try again.", ""
|
109 |
-
continue
|
110 |
-
|
111 |
-
for response in response_stream:
|
112 |
-
if isinstance(response, dict) and "choices" in response:
|
113 |
-
for choice in response["choices"]:
|
114 |
-
if "delta" in choice and "content" in choice["delta"]:
|
115 |
-
chunk = choice["delta"]["content"]
|
116 |
-
full_response += chunk
|
117 |
-
yield full_response, ""
|
118 |
-
else:
|
119 |
-
logging.error(f"Unexpected response format in API call {call + 1}: {response}")
|
120 |
-
|
121 |
-
if full_response:
|
122 |
-
break # If we got a valid response, exit the loop
|
123 |
-
|
124 |
-
except Exception as e:
|
125 |
-
logging.error(f"Error in API call {call + 1}: {str(e)}")
|
126 |
-
if "422 Client Error" in str(e):
|
127 |
-
logging.warning("Received 422 Client Error. Adjusting request parameters.")
|
128 |
-
# You might want to adjust parameters here, e.g., reduce max_tokens
|
129 |
-
yield f"An error occurred during API call {call + 1}. Retrying...", ""
|
130 |
-
|
131 |
-
await asyncio.sleep(1) # 1 second delay between calls
|
132 |
|
133 |
-
|
134 |
-
|
135 |
-
|
136 |
-
|
137 |
-
|
138 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
139 |
|
140 |
if not full_response:
|
141 |
logging.warning("No response generated from the model")
|
@@ -143,18 +131,12 @@ Write a detailed and complete research document that fulfills the following user
|
|
143 |
else:
|
144 |
yield f"{full_response}\n\nSources:\n{source_list_str}", ""
|
145 |
|
146 |
-
|
147 |
-
logging.info(f"User Query: {message}")
|
148 |
-
logging.info(f"Model Used: {model}")
|
149 |
-
logging.info(f"Temperature: {temperature}")
|
150 |
-
logging.info(f"Number of API Calls: {num_calls}")
|
151 |
-
logging.info(f"Use Embeddings: {use_embeddings}")
|
152 |
-
logging.info(f"System Prompt: {system_prompt}")
|
153 |
-
logging.info(f"History: {history}") # Log the history for debugging
|
154 |
-
|
155 |
-
# Convert gradio history to the format expected by get_response_with_search
|
156 |
chat_history = []
|
157 |
-
if history:
|
|
|
|
|
|
|
158 |
for entry in history:
|
159 |
if isinstance(entry, (list, tuple)) and len(entry) == 2:
|
160 |
human, assistant = entry
|
@@ -164,10 +146,20 @@ async def respond(message, system_prompt, history, model, temperature, num_calls
|
|
164 |
elif isinstance(entry, str):
|
165 |
# If it's a string, assume it's a user message
|
166 |
chat_history.append({"role": "user", "content": entry})
|
167 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
168 |
|
169 |
try:
|
170 |
-
full_response = ""
|
171 |
async for main_content, sources in get_response_with_search(
|
172 |
message,
|
173 |
system_prompt,
|
@@ -177,16 +169,8 @@ async def respond(message, system_prompt, history, model, temperature, num_calls
|
|
177 |
num_calls=num_calls,
|
178 |
temperature=temperature
|
179 |
):
|
180 |
-
|
181 |
-
|
182 |
-
yield main_content
|
183 |
-
else:
|
184 |
-
# Otherwise, yield only the new content
|
185 |
-
new_content = main_content[len(full_response):]
|
186 |
-
full_response = main_content
|
187 |
-
yield new_content
|
188 |
-
|
189 |
-
# Yield the sources as a separate message
|
190 |
if sources:
|
191 |
yield f"\n\nSources:\n{sources}"
|
192 |
|
@@ -213,16 +197,8 @@ css = """
|
|
213 |
def create_gradio_interface():
|
214 |
custom_placeholder = "Enter your question here for web search."
|
215 |
|
216 |
-
async def wrapped_respond(*args):
|
217 |
-
try:
|
218 |
-
async for response in respond(*args):
|
219 |
-
yield response
|
220 |
-
except Exception as e:
|
221 |
-
logging.error(f"Error in wrapped_respond: {str(e)}")
|
222 |
-
yield f"An error occurred: {str(e)}"
|
223 |
-
|
224 |
demo = gr.ChatInterface(
|
225 |
-
fn=
|
226 |
additional_inputs_accordion=gr.Accordion(label="⚙️ Parameters", open=True, render=False),
|
227 |
additional_inputs=[
|
228 |
gr.Textbox(value=DEFAULT_SYSTEM_PROMPT, lines=6, label="System Prompt", placeholder="Enter your system prompt here"),
|
@@ -268,4 +244,4 @@ def create_gradio_interface():
|
|
268 |
|
269 |
if __name__ == "__main__":
|
270 |
demo = create_gradio_interface()
|
271 |
-
demo.launch(share=True)
|
|
|
1 |
import os
|
2 |
import logging
|
3 |
import asyncio
|
4 |
+
from typing import AsyncGenerator, Tuple
|
5 |
import gradio as gr
|
6 |
from huggingface_hub import InferenceClient
|
7 |
from langchain.embeddings import HuggingFaceEmbeddings
|
|
|
56 |
logging.info(f"Created vectors for {len(documents)} search results.")
|
57 |
return FAISS.from_documents(documents, embed)
|
58 |
|
59 |
+
def create_context(search_results, use_embeddings, query):
|
60 |
+
if use_embeddings:
|
61 |
+
web_search_database = create_web_search_vectors(search_results)
|
62 |
+
retriever = web_search_database.as_retriever(search_kwargs={"k": 5})
|
63 |
+
relevant_docs = retriever.get_relevant_documents(query)
|
64 |
+
return "\n".join([doc.page_content for doc in relevant_docs])
|
65 |
+
else:
|
66 |
+
return "\n".join([f"{result['title']}\n{result['body']}" for result in search_results])
|
67 |
+
|
68 |
+
async def get_response_with_search(query: str, system_prompt: str, model: str, use_embeddings: bool, history=None, num_calls: int = 3, temperature: float = 0.2) -> AsyncGenerator[Tuple[str, str], None]:
|
69 |
search_results = duckduckgo_search(query)
|
70 |
|
71 |
if not search_results:
|
|
|
76 |
sources = [result['href'] for result in search_results if 'href' in result]
|
77 |
source_list_str = "\n".join(sources)
|
78 |
|
79 |
+
context = create_context(search_results, use_embeddings, query)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
80 |
logging.info(f"Context created for query: {query}")
|
81 |
|
82 |
user_message = f"""Using the following context from web search results:
|
|
|
84 |
|
85 |
Write a detailed and complete research document that fulfills the following user request: '{query}'."""
|
86 |
|
|
|
|
|
|
|
87 |
messages = [
|
88 |
{"role": "system", "content": system_prompt},
|
89 |
{"role": "user", "content": user_message}
|
|
|
92 |
if history:
|
93 |
messages = history + messages
|
94 |
|
95 |
+
client = InferenceClient(model, token=huggingface_token)
|
96 |
+
full_response = ""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
97 |
|
98 |
+
for call in range(num_calls):
|
99 |
+
try:
|
100 |
+
response = await asyncio.to_thread(
|
101 |
+
client.chat_completion,
|
102 |
+
messages=messages,
|
103 |
+
max_tokens=6000,
|
104 |
+
temperature=temperature,
|
105 |
+
top_p=0.8,
|
106 |
+
)
|
107 |
+
|
108 |
+
if response is None or not isinstance(response, dict) or 'choices' not in response:
|
109 |
+
logging.error(f"API call {call + 1} returned an invalid response: {response}")
|
110 |
+
if call == num_calls - 1:
|
111 |
+
yield "The API returned an invalid response. Please try again later.", ""
|
112 |
+
continue
|
113 |
+
|
114 |
+
new_content = response['choices'][0]['message']['content']
|
115 |
+
full_response += new_content
|
116 |
+
yield full_response, ""
|
117 |
+
|
118 |
+
if full_response:
|
119 |
+
break # If we got a valid response, exit the loop
|
120 |
+
|
121 |
+
except Exception as e:
|
122 |
+
logging.error(f"Error in API call {call + 1}: {str(e)}")
|
123 |
+
if call == num_calls - 1:
|
124 |
+
yield f"An error occurred during API calls: {str(e)}. Please try again later.", ""
|
125 |
+
|
126 |
+
await asyncio.sleep(1) # 1 second delay between calls
|
127 |
|
128 |
if not full_response:
|
129 |
logging.warning("No response generated from the model")
|
|
|
131 |
else:
|
132 |
yield f"{full_response}\n\nSources:\n{source_list_str}", ""
|
133 |
|
134 |
+
def process_history(history):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
135 |
chat_history = []
|
136 |
+
if isinstance(history, str):
|
137 |
+
# If history is a string (like the system prompt), add it as a system message
|
138 |
+
chat_history.append({"role": "system", "content": history})
|
139 |
+
elif isinstance(history, list):
|
140 |
for entry in history:
|
141 |
if isinstance(entry, (list, tuple)) and len(entry) == 2:
|
142 |
human, assistant = entry
|
|
|
146 |
elif isinstance(entry, str):
|
147 |
# If it's a string, assume it's a user message
|
148 |
chat_history.append({"role": "user", "content": entry})
|
149 |
+
return chat_history
|
150 |
+
|
151 |
+
async def respond(message, system_prompt, history, model, temperature, num_calls, use_embeddings):
|
152 |
+
logging.info(f"User Query: {message}")
|
153 |
+
logging.info(f"Model Used: {model}")
|
154 |
+
logging.info(f"Temperature: {temperature}")
|
155 |
+
logging.info(f"Number of API Calls: {num_calls}")
|
156 |
+
logging.info(f"Use Embeddings: {use_embeddings}")
|
157 |
+
logging.info(f"System Prompt: {system_prompt}")
|
158 |
+
logging.info(f"History: {history}")
|
159 |
+
|
160 |
+
chat_history = process_history(history)
|
161 |
|
162 |
try:
|
|
|
163 |
async for main_content, sources in get_response_with_search(
|
164 |
message,
|
165 |
system_prompt,
|
|
|
169 |
num_calls=num_calls,
|
170 |
temperature=temperature
|
171 |
):
|
172 |
+
yield main_content
|
173 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
174 |
if sources:
|
175 |
yield f"\n\nSources:\n{sources}"
|
176 |
|
|
|
197 |
def create_gradio_interface():
|
198 |
custom_placeholder = "Enter your question here for web search."
|
199 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
200 |
demo = gr.ChatInterface(
|
201 |
+
fn=respond,
|
202 |
additional_inputs_accordion=gr.Accordion(label="⚙️ Parameters", open=True, render=False),
|
203 |
additional_inputs=[
|
204 |
gr.Textbox(value=DEFAULT_SYSTEM_PROMPT, lines=6, label="System Prompt", placeholder="Enter your system prompt here"),
|
|
|
244 |
|
245 |
if __name__ == "__main__":
|
246 |
demo = create_gradio_interface()
|
247 |
+
demo.launch(share=True)
|