Spaces:
Running
Running
abubasith86
commited on
Commit
·
1a8ed74
1
Parent(s):
a5eeaad
revoked
Browse files- app.py +10 -60
- requirements.txt +1 -3
app.py
CHANGED
|
@@ -1,28 +1,10 @@
|
|
| 1 |
import gradio as gr
|
| 2 |
from huggingface_hub import InferenceClient
|
| 3 |
-
import pymupdf
|
| 4 |
-
from duckduckgo_search import DDGS
|
| 5 |
|
| 6 |
"""
|
| 7 |
For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
|
| 8 |
"""
|
| 9 |
-
client = InferenceClient("
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
# PDF Parsing
|
| 13 |
-
def extract_text_from_pdf(pdf_file):
|
| 14 |
-
doc = pymupdf.open(pdf_file)
|
| 15 |
-
text = " ".join([page.get_textpage().extractTEXT() for page in doc])
|
| 16 |
-
return text
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
# Web search fallback
|
| 20 |
-
def search_web(query):
|
| 21 |
-
with DDGS() as ddgs:
|
| 22 |
-
results = ddgs.text(query)
|
| 23 |
-
if results:
|
| 24 |
-
return results[0]["body"]
|
| 25 |
-
return "No relevant results found on the web."
|
| 26 |
|
| 27 |
|
| 28 |
def respond(
|
|
@@ -32,51 +14,19 @@ def respond(
|
|
| 32 |
max_tokens,
|
| 33 |
temperature,
|
| 34 |
top_p,
|
| 35 |
-
pdf_file=None,
|
| 36 |
):
|
| 37 |
-
|
| 38 |
-
"latest",
|
| 39 |
-
"today",
|
| 40 |
-
"current",
|
| 41 |
-
"now",
|
| 42 |
-
"recent",
|
| 43 |
-
"news",
|
| 44 |
-
"update",
|
| 45 |
-
"price",
|
| 46 |
-
"who won",
|
| 47 |
-
"what happened",
|
| 48 |
-
]
|
| 49 |
-
message_lower = message.lower()
|
| 50 |
-
|
| 51 |
-
# Check if this is a recent/live query
|
| 52 |
-
if any(kw in message_lower for kw in recent_keywords):
|
| 53 |
-
web_result = search_web(message)
|
| 54 |
-
if web_result:
|
| 55 |
-
yield f"[Answer from Web Search]\n{web_result}"
|
| 56 |
-
return
|
| 57 |
-
|
| 58 |
-
# Check if PDF was uploaded
|
| 59 |
-
context_text = ""
|
| 60 |
-
if pdf_file is not None:
|
| 61 |
-
try:
|
| 62 |
-
context_text = extract_text_from_pdf(pdf_file.name)
|
| 63 |
-
except Exception as e:
|
| 64 |
-
yield f"[Error reading PDF: {str(e)}]"
|
| 65 |
|
| 66 |
-
|
| 67 |
-
|
| 68 |
-
|
|
|
|
|
|
|
| 69 |
|
| 70 |
-
# Fallback to LLM (Zephyr)
|
| 71 |
-
messages = [{"role": "system", "content": system_message}]
|
| 72 |
-
for user, assistant in history:
|
| 73 |
-
if user:
|
| 74 |
-
messages.append({"role": "user", "content": user})
|
| 75 |
-
if assistant:
|
| 76 |
-
messages.append({"role": "assistant", "content": assistant})
|
| 77 |
messages.append({"role": "user", "content": message})
|
| 78 |
|
| 79 |
response = ""
|
|
|
|
| 80 |
for message in client.chat_completion(
|
| 81 |
messages,
|
| 82 |
max_tokens=max_tokens,
|
|
@@ -85,7 +35,8 @@ def respond(
|
|
| 85 |
top_p=top_p,
|
| 86 |
):
|
| 87 |
token = message.choices[0].delta.content
|
| 88 |
-
|
|
|
|
| 89 |
yield response
|
| 90 |
|
| 91 |
|
|
@@ -105,7 +56,6 @@ demo = gr.ChatInterface(
|
|
| 105 |
step=0.05,
|
| 106 |
label="Top-p (nucleus sampling)",
|
| 107 |
),
|
| 108 |
-
gr.File(label="Upload a PDF", file_types=[".pdf"]),
|
| 109 |
],
|
| 110 |
)
|
| 111 |
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
from huggingface_hub import InferenceClient
|
|
|
|
|
|
|
| 3 |
|
| 4 |
"""
|
| 5 |
For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
|
| 6 |
"""
|
| 7 |
+
client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
|
| 9 |
|
| 10 |
def respond(
|
|
|
|
| 14 |
max_tokens,
|
| 15 |
temperature,
|
| 16 |
top_p,
|
|
|
|
| 17 |
):
|
| 18 |
+
messages = [{"role": "system", "content": system_message}]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
+
for val in history:
|
| 21 |
+
if val[0]:
|
| 22 |
+
messages.append({"role": "user", "content": val[0]})
|
| 23 |
+
if val[1]:
|
| 24 |
+
messages.append({"role": "assistant", "content": val[1]})
|
| 25 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
messages.append({"role": "user", "content": message})
|
| 27 |
|
| 28 |
response = ""
|
| 29 |
+
|
| 30 |
for message in client.chat_completion(
|
| 31 |
messages,
|
| 32 |
max_tokens=max_tokens,
|
|
|
|
| 35 |
top_p=top_p,
|
| 36 |
):
|
| 37 |
token = message.choices[0].delta.content
|
| 38 |
+
|
| 39 |
+
response += token
|
| 40 |
yield response
|
| 41 |
|
| 42 |
|
|
|
|
| 56 |
step=0.05,
|
| 57 |
label="Top-p (nucleus sampling)",
|
| 58 |
),
|
|
|
|
| 59 |
],
|
| 60 |
)
|
| 61 |
|
requirements.txt
CHANGED
|
@@ -1,3 +1 @@
|
|
| 1 |
-
huggingface_hub==0.25.2
|
| 2 |
-
duckduckgo_search
|
| 3 |
-
pymupdf
|
|
|
|
| 1 |
+
huggingface_hub==0.25.2
|
|
|
|
|
|