Spaces:
Sleeping
Sleeping
feat: :lock: Implement API Key authentication and CORS configuration for enhanced security. Don`t returns thread ID
Browse files
app.py
CHANGED
|
@@ -4,6 +4,9 @@ from transformers import AutoModelForCausalLM, AutoTokenizer
|
|
| 4 |
import torch
|
| 5 |
from functools import partial
|
| 6 |
from fastapi.responses import JSONResponse
|
|
|
|
|
|
|
|
|
|
| 7 |
|
| 8 |
from langchain_core.messages import HumanMessage, AIMessage
|
| 9 |
from langgraph.checkpoint.memory import MemorySaver
|
|
@@ -13,6 +16,19 @@ import os
|
|
| 13 |
from dotenv import load_dotenv
|
| 14 |
load_dotenv()
|
| 15 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
# Initialize the model and tokenizer
|
| 17 |
print("Loading model and tokenizer...")
|
| 18 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
|
@@ -120,7 +136,45 @@ class SummaryRequest(BaseModel):
|
|
| 120 |
max_length: int = 200
|
| 121 |
|
| 122 |
# Create the FastAPI application
|
| 123 |
-
app = FastAPI(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 124 |
|
| 125 |
# Add general exception handler
|
| 126 |
@app.exception_handler(Exception)
|
|
@@ -138,7 +192,10 @@ async def api_home():
|
|
| 138 |
|
| 139 |
# Generate endpoint
|
| 140 |
@app.post("/generate")
|
| 141 |
-
async def generate(
|
|
|
|
|
|
|
|
|
|
| 142 |
"""
|
| 143 |
Endpoint to generate text using the language model
|
| 144 |
|
|
@@ -147,9 +204,10 @@ async def generate(request: QueryRequest):
|
|
| 147 |
query: str
|
| 148 |
thread_id: str = "default"
|
| 149 |
system_prompt: str = DEFAULT_SYSTEM_PROMPT
|
|
|
|
| 150 |
|
| 151 |
Returns:
|
| 152 |
-
dict: A dictionary containing the generated text
|
| 153 |
"""
|
| 154 |
try:
|
| 155 |
# Configure the thread ID
|
|
@@ -175,21 +233,22 @@ async def generate(request: QueryRequest):
|
|
| 175 |
response = output["messages"][-1].content
|
| 176 |
|
| 177 |
return {
|
| 178 |
-
"generated_text": response
|
| 179 |
-
"thread_id": request.thread_id
|
| 180 |
}
|
| 181 |
except Exception as e:
|
| 182 |
return JSONResponse(
|
| 183 |
status_code=500,
|
| 184 |
content={
|
| 185 |
"error": f"Error generando texto: {str(e)}",
|
| 186 |
-
"type": type(e).__name__
|
| 187 |
-
"thread_id": request.thread_id
|
| 188 |
}
|
| 189 |
)
|
| 190 |
|
| 191 |
@app.post("/summarize")
|
| 192 |
-
async def summarize(
|
|
|
|
|
|
|
|
|
|
| 193 |
"""
|
| 194 |
Endpoint to generate a summary using the language model
|
| 195 |
|
|
@@ -198,9 +257,10 @@ async def summarize(request: SummaryRequest):
|
|
| 198 |
text: str - The text to summarize
|
| 199 |
thread_id: str = "default"
|
| 200 |
max_length: int = 200 - Maximum summary length
|
|
|
|
| 201 |
|
| 202 |
Returns:
|
| 203 |
-
dict: A dictionary containing the summary
|
| 204 |
"""
|
| 205 |
try:
|
| 206 |
# Configure the thread ID
|
|
@@ -228,16 +288,14 @@ async def summarize(request: SummaryRequest):
|
|
| 228 |
response = output["messages"][-1].content
|
| 229 |
|
| 230 |
return {
|
| 231 |
-
"summary": response
|
| 232 |
-
"thread_id": request.thread_id
|
| 233 |
}
|
| 234 |
except Exception as e:
|
| 235 |
return JSONResponse(
|
| 236 |
status_code=500,
|
| 237 |
content={
|
| 238 |
"error": f"Error generando resumen: {str(e)}",
|
| 239 |
-
"type": type(e).__name__
|
| 240 |
-
"thread_id": request.thread_id
|
| 241 |
}
|
| 242 |
)
|
| 243 |
|
|
|
|
| 4 |
import torch
|
| 5 |
from functools import partial
|
| 6 |
from fastapi.responses import JSONResponse
|
| 7 |
+
from fastapi import Security, Depends
|
| 8 |
+
from fastapi.security.api_key import APIKeyHeader, APIKey
|
| 9 |
+
from fastapi.middleware.cors import CORSMiddleware
|
| 10 |
|
| 11 |
from langchain_core.messages import HumanMessage, AIMessage
|
| 12 |
from langgraph.checkpoint.memory import MemorySaver
|
|
|
|
| 16 |
from dotenv import load_dotenv
|
| 17 |
load_dotenv()
|
| 18 |
|
| 19 |
+
# Configuración de API Key
|
| 20 |
+
API_KEY_NAME = "X-API-Key"
|
| 21 |
+
API_KEY = os.getenv("API_KEY")
|
| 22 |
+
api_key_header = APIKeyHeader(name=API_KEY_NAME, auto_error=False)
|
| 23 |
+
|
| 24 |
+
async def get_api_key(api_key_header: str = Security(api_key_header)):
|
| 25 |
+
if api_key_header == API_KEY:
|
| 26 |
+
return api_key_header
|
| 27 |
+
raise HTTPException(
|
| 28 |
+
status_code=403,
|
| 29 |
+
detail="Could not validate API KEY"
|
| 30 |
+
)
|
| 31 |
+
|
| 32 |
# Initialize the model and tokenizer
|
| 33 |
print("Loading model and tokenizer...")
|
| 34 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
|
|
|
| 136 |
max_length: int = 200
|
| 137 |
|
| 138 |
# Create the FastAPI application
|
| 139 |
+
app = FastAPI(
|
| 140 |
+
title="LangChain FastAPI",
|
| 141 |
+
description="API to generate text using LangChain and LangGraph - Máximo Fernández Núñez IriusRisk test challenge",
|
| 142 |
+
version="1.0.0",
|
| 143 |
+
openapi_tags=[
|
| 144 |
+
{
|
| 145 |
+
"name": "Authentication",
|
| 146 |
+
"description": "Endpoints require API Key authentication via X-API-Key header"
|
| 147 |
+
}
|
| 148 |
+
]
|
| 149 |
+
)
|
| 150 |
+
|
| 151 |
+
# Configure the security scheme in the OpenAPI documentation
|
| 152 |
+
app.openapi_tags = [
|
| 153 |
+
{"name": "Authentication", "description": "Protected endpoints that require API Key"}
|
| 154 |
+
]
|
| 155 |
+
|
| 156 |
+
# Import and configure CORS
|
| 157 |
+
app.add_middleware(
|
| 158 |
+
CORSMiddleware,
|
| 159 |
+
allow_origins=["*"],
|
| 160 |
+
allow_credentials=True,
|
| 161 |
+
allow_methods=["*"],
|
| 162 |
+
allow_headers=["*"],
|
| 163 |
+
)
|
| 164 |
+
|
| 165 |
+
# Configure the security scheme
|
| 166 |
+
app.openapi_components = {
|
| 167 |
+
"securitySchemes": {
|
| 168 |
+
"api_key": {
|
| 169 |
+
"type": "apiKey",
|
| 170 |
+
"name": API_KEY_NAME,
|
| 171 |
+
"in": "header",
|
| 172 |
+
"description": "Enter your API key"
|
| 173 |
+
}
|
| 174 |
+
}
|
| 175 |
+
}
|
| 176 |
+
|
| 177 |
+
app.openapi_security = [{"api_key": []}]
|
| 178 |
|
| 179 |
# Add general exception handler
|
| 180 |
@app.exception_handler(Exception)
|
|
|
|
| 192 |
|
| 193 |
# Generate endpoint
|
| 194 |
@app.post("/generate")
|
| 195 |
+
async def generate(
|
| 196 |
+
request: QueryRequest,
|
| 197 |
+
api_key: APIKey = Depends(get_api_key)
|
| 198 |
+
):
|
| 199 |
"""
|
| 200 |
Endpoint to generate text using the language model
|
| 201 |
|
|
|
|
| 204 |
query: str
|
| 205 |
thread_id: str = "default"
|
| 206 |
system_prompt: str = DEFAULT_SYSTEM_PROMPT
|
| 207 |
+
api_key: APIKey - API key for authentication
|
| 208 |
|
| 209 |
Returns:
|
| 210 |
+
dict: A dictionary containing the generated text
|
| 211 |
"""
|
| 212 |
try:
|
| 213 |
# Configure the thread ID
|
|
|
|
| 233 |
response = output["messages"][-1].content
|
| 234 |
|
| 235 |
return {
|
| 236 |
+
"generated_text": response
|
|
|
|
| 237 |
}
|
| 238 |
except Exception as e:
|
| 239 |
return JSONResponse(
|
| 240 |
status_code=500,
|
| 241 |
content={
|
| 242 |
"error": f"Error generando texto: {str(e)}",
|
| 243 |
+
"type": type(e).__name__
|
|
|
|
| 244 |
}
|
| 245 |
)
|
| 246 |
|
| 247 |
@app.post("/summarize")
|
| 248 |
+
async def summarize(
|
| 249 |
+
request: SummaryRequest,
|
| 250 |
+
api_key: APIKey = Depends(get_api_key)
|
| 251 |
+
):
|
| 252 |
"""
|
| 253 |
Endpoint to generate a summary using the language model
|
| 254 |
|
|
|
|
| 257 |
text: str - The text to summarize
|
| 258 |
thread_id: str = "default"
|
| 259 |
max_length: int = 200 - Maximum summary length
|
| 260 |
+
api_key: APIKey - API key for authentication
|
| 261 |
|
| 262 |
Returns:
|
| 263 |
+
dict: A dictionary containing the summary
|
| 264 |
"""
|
| 265 |
try:
|
| 266 |
# Configure the thread ID
|
|
|
|
| 288 |
response = output["messages"][-1].content
|
| 289 |
|
| 290 |
return {
|
| 291 |
+
"summary": response
|
|
|
|
| 292 |
}
|
| 293 |
except Exception as e:
|
| 294 |
return JSONResponse(
|
| 295 |
status_code=500,
|
| 296 |
content={
|
| 297 |
"error": f"Error generando resumen: {str(e)}",
|
| 298 |
+
"type": type(e).__name__
|
|
|
|
| 299 |
}
|
| 300 |
)
|
| 301 |
|