File size: 27,091 Bytes
8cb6e00 7662cb9 b6c7c7c 8cb6e00 9ea4e14 8cb6e00 c0f33c6 0b01c12 8cb6e00 3c9eeaf 8cb6e00 4fbcf68 9ea4e14 8a7f2d8 0c663c0 8cb6e00 b6c7c7c 5c0b5cd b6c7c7c 5c0b5cd ec0e22d be86181 ae037a5 8cb6e00 7f09169 8cb6e00 d784ff5 4ee7a1a 8cb6e00 c9feee8 8cb6e00 b6c7c7c 0b01c12 b6c7c7c 8cb6e00 b6c7c7c 8cb6e00 1b538b8 8cb6e00 aa0bf91 d784ff5 aa0bf91 e40819a aa0bf91 8cb6e00 b6c7c7c 8cb6e00 c9feee8 8cb6e00 b6c7c7c 0b01c12 8cb6e00 b6c7c7c 8cb6e00 71bda1d 9ea4e14 71bda1d c0f33c6 71bda1d f9467b5 71bda1d c0f33c6 f6fc26e 71bda1d c0f33c6 3dd8d6c c0f33c6 f9467b5 71bda1d f9467b5 71bda1d 3dd8d6c 9ea4e14 8cb6e00 d784ff5 8cb6e00 18e298f 8cb6e00 92b1ad1 7a95386 8cb6e00 7a95386 8cb6e00 9395959 8cb6e00 ae28beb 8cb6e00 7a95386 92b1ad1 7a95386 a20826a 7a95386 8cb6e00 c9feee8 8cb6e00 1486ce7 48e6960 1486ce7 d784ff5 830b4bc 0dbfbd4 12c278e 0dbfbd4 4fbcf68 0dbfbd4 d67a459 8cb6e00 d7d1d4e 8cb6e00 f8d95b7 0c663c0 0dbfbd4 d7d1d4e 01e823e 1211e2b 0dbfbd4 01e823e d7d1d4e 8cb6e00 b6c7c7c 8cb6e00 a0d14fd 8cb6e00 a0d14fd fbda383 a0d14fd 8cb6e00 a0d14fd 8cb6e00 a0d14fd 8cb6e00 b8d0141 8cb6e00 a0d14fd 8cb6e00 a0d14fd 8cb6e00 a0d14fd 60c50d7 3dd8d6c b8d0141 a0d14fd 8cb6e00 a0d14fd 8cb6e00 a0d14fd 8cb6e00 a0d14fd 18e298f a0d14fd 8cb6e00 a0d14fd 8cb6e00 a0d14fd 8cb6e00 a0d14fd 92b1ad1 a0d14fd a20826a a0d14fd b6c7c7c a0d14fd 8cb6e00 a0d14fd 8cb6e00 b8d0141 a0d14fd eab6a7c b8d0141 2ca501d 5022fdb 66fea52 eab6a7c b8d0141 9395959 b8d0141 ae28beb b8d0141 6efb035 b8d0141 b6c7c7c b8d0141 7a95386 92b1ad1 7a95386 a20826a 7a95386 b8d0141 7a95386 b8d0141 6733cba 48e6960 830b4bc d784ff5 3c9eeaf 0dbfbd4 5f2bc85 0dbfbd4 67a965a eab6a7c 0dbfbd4 4fbcf68 0dbfbd4 4fbcf68 0dbfbd4 1486ce7 b8d0141 80a7dbc 1486ce7 0dbfbd4 0c663c0 f8d95b7 0c663c0 0dbfbd4 b8d0141 1486ce7 0c663c0 0dbfbd4 b8d0141 b6c7c7c 4167849 b8d0141 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 |
# Import necessary modules
from concurrent.futures import ProcessPoolExecutor
import logging
import os
import asyncio
import threading
import uuid
from fastapi import FastAPI, HTTPException, Header
from fastapi.encoders import jsonable_encoder
from typing import Dict, List, Literal, Optional
from fastapi.responses import FileResponse
import numpy as np
import pandas as pd
from pandasai import SmartDataframe
from langchain_groq.chat_models import ChatGroq
from dotenv import load_dotenv
from pydantic import BaseModel, Field, ValidationError
from csv_service import clean_data, extract_chart_filenames, generate_csv_data, get_csv_basic_info
from urllib.parse import unquote
from langchain_groq import ChatGroq
import pandas as pd
from langchain_experimental.tools import PythonAstREPLTool
from langchain_experimental.agents import create_pandas_dataframe_agent
import numpy as np
import matplotlib.pyplot as plt
import matplotlib
import seaborn as sns
from gemini_report_generator import generate_csv_report
from intitial_q_handler import if_initial_chart_question, if_initial_chat_question
from orchestrator_agent import csv_orchestrator_chat
from python_code_executor_service import CsvChatResult, PythonExecutor
from supabase_service import upload_file_to_supabase
from cerebras_csv_agent import query_csv_agent
from util_service import _prompt_generator, process_answer
from fastapi.middleware.cors import CORSMiddleware
import matplotlib
matplotlib.use('Agg')
# Initialize FastAPI app
app = FastAPI()
# Set up logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
# Initialize the ProcessPoolExecutor
max_cpus = os.cpu_count()
logger.info(f"Max CPUs: {max_cpus}")
# Ensure the cache directory exists
os.makedirs("/app/cache", exist_ok=True)
os.makedirs("/app", exist_ok=True)
open("/app/pandasai.log", "a").close() # Create the file if it doesn't exist
# Ensure the generated_charts directory exists
os.makedirs("/app/generated_charts", exist_ok=True)
load_dotenv()
image_file_path = os.getenv("IMAGE_FILE_PATH")
image_not_found = os.getenv("IMAGE_NOT_FOUND")
allowed_hosts = os.getenv("ALLOWED_HOSTS", "").split(",")
app.add_middleware(
CORSMiddleware,
allow_origins=allowed_hosts,
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# Load environment variables
groq_api_keys = os.getenv("GROQ_API_KEYS").split(",")
model_name = os.getenv("GROQ_LLM_MODEL")
class CsvUrlRequest(BaseModel):
csv_url: str
class ImageRequest(BaseModel):
image_path: str
chat_id: str
class FileProps(BaseModel):
fileName: str
filePath: str
fileType: str # 'csv' | 'image'
class Files(BaseModel):
csv_files: List[FileProps]
image_files: List[FileProps]
class FileBoxProps(BaseModel):
files: Files
dummy_response = FileBoxProps(
files=Files(
csv_files=[
FileProps(
fileName="sales_data.csv",
filePath="/downloads/sales_data.csv",
fileType="csv"
),
FileProps(
fileName="customer_data.csv",
filePath="/downloads/customer_data.csv",
fileType="csv"
)
],
image_files=[
FileProps(
fileName="chart.png",
filePath="/downloads/chart.png",
fileType="image"
),
FileProps(
fileName="graph.jpg",
filePath="/downloads/graph.jpg",
fileType="image"
)
]
)
)
# Thread-safe key management for groq_chat
current_groq_key_index = 0
current_groq_key_lock = threading.Lock()
# Thread-safe key management for langchain_csv_chat
current_langchain_key_index = 0
current_langchain_key_lock = threading.Lock()
# PING CHECK
@app.get("/ping")
async def root():
return {"message": "Pong !!"}
# BASIC KNOWLEDGE BASED ON CSV
# Remove trailing slash from the URL otherwise it will redirect to GET method
@app.post("/api/basic_csv_data")
async def basic_csv_data(request: CsvUrlRequest):
try:
decoded_url = unquote(request.csv_url)
logger.info(f"Fetching CSV data from URL: {decoded_url}")
# csv_data = await get_csv_basic_info(decoded_url)
# Run the synchronous function in a thread pool executor
loop = asyncio.get_running_loop()
csv_data = await loop.run_in_executor(
process_executor, get_csv_basic_info, decoded_url
)
logger.info(f"CSV data fetched successfully: {csv_data}")
return {"data": csv_data}
except Exception as e:
logger.error(f"Error while fetching CSV data: {e}")
raise HTTPException(status_code=400, detail=f"Failed to retrieve CSV data: {str(e)}")
# GET THE CHART FROM A SPECIFIC FILE PATH
@app.post("/api/get-chart")
async def get_image(request: ImageRequest, authorization: str = Header(None)):
if not authorization:
raise HTTPException(status_code=401, detail="Authorization header missing")
if not authorization.startswith("Bearer "):
raise HTTPException(status_code=401, detail="Invalid authorization header format")
token = authorization.split(" ")[1]
if not token:
raise HTTPException(status_code=401, detail="Token missing")
if token != os.getenv("AUTH_TOKEN"):
raise HTTPException(status_code=403, detail="Invalid token")
try:
logger.info("Groq Chat created a chat for the user query...")
image_file_path = request.image_path
unique_file_name =f'{str(uuid.uuid4())}.png'
logger.info("Uploading the chart to supabase...")
image_public_url = await upload_file_to_supabase(f"{image_file_path}", unique_file_name, chat_id=request.chat_id)
logger.info("Image uploaded to Supabase and Image URL is... ", {image_public_url})
os.remove(image_file_path)
return {"image_url": image_public_url}
# return FileResponse(image_file_path, media_type="image/png")
except Exception as e:
logger.error(f"Error: {e}")
return {"answer": "error"}
# GET CSV DATA FOR GENERATING THE TABLE
@app.post("/api/csv_data")
async def get_csv_data(request: CsvUrlRequest):
try:
decoded_url = unquote(request.csv_url)
logger.info(f"Fetching CSV data from URL: {decoded_url}")
# csv_data = await generate_csv_data(decoded_url)
loop = asyncio.get_running_loop()
csv_data = await loop.run_in_executor(
process_executor, generate_csv_data, decoded_url
)
return csv_data
except Exception as e:
logger.error(f"Error while fetching CSV data: {e}")
raise HTTPException(status_code=400, detail=f"Failed to retrieve CSV data: {str(e)}")
# EXECUTE THE PYTHON CODE
class ExecutionRequest(BaseModel):
chat_id: str = Field(..., alias="chat_id")
csv_url: str = Field(..., alias="csv_url")
codeExecutionPayload: CsvChatResult
@app.post("/api/code_execution_csv")
async def code_execution_csv(
request_data: ExecutionRequest, # Change from ExecutionRequest to dict to see raw input
authorization: Optional[str] = Header(None)
):
# Auth check remains the same
expected_token = os.environ.get("AUTH_TOKEN")
if not authorization or not expected_token or authorization.replace("Bearer ", "") != expected_token:
raise HTTPException(status_code=401, detail="Unauthorized")
try:
# First log the incoming request data
logger.info("Incoming request data:", request_data)
# Rest of your processing logic...
decoded_url = unquote(request_data.csv_url)
df = clean_data(decoded_url)
executor = PythonExecutor(df)
formatted_output = await executor.process_response(request_data.codeExecutionPayload, request_data.chat_id)
return {"answer": formatted_output}
except Exception as e:
logger.info("Processing error:", str(e))
return {"error": "Failed to execute request", "message": str(e)}
# CHAT CODING STARTS FROM HERE
# Modified groq_chat function with thread-safe key rotation
def groq_chat(csv_url: str, question: str):
global current_groq_key_index, current_groq_key_lock
while True:
with current_groq_key_lock:
if current_groq_key_index >= len(groq_api_keys):
return {"error": "All API keys exhausted."}
current_api_key = groq_api_keys[current_groq_key_index]
try:
data = clean_data(csv_url)
llm = ChatGroq(model=model_name, api_key=current_api_key)
# Generate unique filename using UUID
chart_filename = f"chart_{uuid.uuid4()}.png"
chart_path = os.path.join("generated_charts", chart_filename)
# Configure SmartDataframe with chart settings
df = SmartDataframe(
data,
config={
'llm': llm,
'save_charts': True, # Enable chart saving
'open_charts': False,
'save_charts_path': os.path.dirname(chart_path), # Directory to save
'custom_chart_filename': chart_filename, # Unique filename
'enable_cache': False
}
)
answer = df.chat(question)
# Process different response types
if isinstance(answer, pd.DataFrame):
processed = answer.apply(handle_out_of_range_float).to_dict(orient="records")
elif isinstance(answer, pd.Series):
processed = answer.apply(handle_out_of_range_float).to_dict()
elif isinstance(answer, list):
processed = [handle_out_of_range_float(item) for item in answer]
elif isinstance(answer, dict):
processed = {k: handle_out_of_range_float(v) for k, v in answer.items()}
else:
processed = {"answer": str(handle_out_of_range_float(answer))}
return processed
except Exception as e:
error_message = str(e)
if error_message != "":
logger.warning("Rate limit exceeded. Switching to next API key.")
with current_groq_key_lock:
current_groq_key_index += 1
if current_groq_key_index >= len(groq_api_keys):
return {"error": "All API keys exhausted."}
else:
logger.error("Error in groq_chat: %s", e)
return {"error": error_message}
# Modified langchain_csv_chat with thread-safe key rotation
def langchain_csv_chat(csv_url: str, question: str, chart_required: bool):
global current_langchain_key_index, current_langchain_key_lock
data = clean_data(csv_url)
attempts = 0
while attempts < len(groq_api_keys):
with current_langchain_key_lock:
if current_langchain_key_index >= len(groq_api_keys):
current_langchain_key_index = 0
api_key = groq_api_keys[current_langchain_key_index]
current_key = current_langchain_key_index
current_langchain_key_index += 1
attempts += 1
try:
llm = ChatGroq(model=model_name, api_key=api_key)
tool = PythonAstREPLTool(locals={
"df": data,
"pd": pd,
"np": np,
"plt": plt,
"sns": sns,
"matplotlib": matplotlib
})
agent = create_pandas_dataframe_agent(
llm,
data,
agent_type="tool-calling",
verbose=True,
allow_dangerous_code=True,
extra_tools=[tool],
return_intermediate_steps=True
)
prompt = _prompt_generator(question, chart_required, csv_url)
result = agent.invoke({"input": prompt})
return result.get("output")
except Exception as e:
error_message = str(e)
# if "429" in error_message:
if error_message != "":
with current_langchain_chart_lock:
current_langchain_chart_key_index = (current_langchain_chart_key_index + 1)
logger.warning(f"Rate limit exceeded. Switching to next API key: {groq_api_keys[current_langchain_chart_key_index]}")
else:
logger.error(f"Error with API key {api_key}: {error_message}")
return {"error": error_message}
return {"error": "All API keys exhausted"}
# Async endpoint with non-blocking execution
@app.post("/api/csv-chat")
async def csv_chat(request: Dict, authorization: str = Header(None)):
# Authorization checks
if not authorization or not authorization.startswith("Bearer "):
raise HTTPException(status_code=401, detail="Invalid authorization")
token = authorization.split(" ")[1]
if token != os.getenv("AUTH_TOKEN"):
raise HTTPException(status_code=403, detail="Invalid token")
try:
query = request.get("query")
csv_url = request.get("csv_url")
decoded_url = unquote(csv_url)
detailed_answer = request.get("detailed_answer")
conversation_history = request.get("conversation_history", [])
generate_report = request.get("generate_report")
chat_id = request.get("chat_id")
if generate_report is True:
report_files = await generate_csv_report(csv_url, query, chat_id, conversation_history)
if report_files is not None:
return {"answer": jsonable_encoder(report_files)}
if if_initial_chat_question(query):
answer = await asyncio.to_thread(
langchain_csv_chat, decoded_url, query, False
)
logger.info("langchain_answer:", answer)
return {"answer": jsonable_encoder(answer)}
# Orchestrate the execution
if detailed_answer is True:
orchestrator_answer = await asyncio.to_thread(
csv_orchestrator_chat, decoded_url, query, conversation_history, chat_id
)
if orchestrator_answer is not None:
return {"answer": jsonable_encoder(orchestrator_answer)}
# Process with groq_chat first
# groq_answer = await asyncio.to_thread(groq_chat, decoded_url, query)
# logger.info("groq_answer:", groq_answer)
result = await query_csv_agent(decoded_url, query, chat_id)
logger.info("cerebras csv answer == >", result)
if result is not None or result == "":
return {"answer": result}
# if process_answer(groq_answer) == "Empty response received.":
# return {"answer": "Sorry, I couldn't find relevant data..."}
# if process_answer(groq_answer):
lang_answer = await asyncio.to_thread(
langchain_csv_chat, decoded_url, query, False
)
if process_answer(lang_answer):
return {"answer": "error"}
return {"answer": jsonable_encoder(lang_answer)}
# return {"answer": jsonable_encoder(groq_answer)}
except Exception as e:
logger.error(f"Error processing request: {str(e)}")
return {"answer": "error"}
def handle_out_of_range_float(value):
if isinstance(value, float):
if np.isnan(value):
return None
elif np.isinf(value):
return "Infinity"
return value
# CHART CODING STARTS FROM HERE
instructions = """
- Please ensure that each value is clearly visible, You may need to adjust the font size, rotate the labels, or use truncation to improve readability (if needed).
- For multiple charts, put all of them in a single file.
- Use colorblind-friendly palette
- Read above instructions and follow them.
"""
# Thread-safe configuration for chart endpoints
current_groq_chart_key_index = 0
current_groq_chart_lock = threading.Lock()
# current_langchain_chart_key_index = 0
# current_langchain_chart_lock = threading.Lock()
def model():
global current_groq_chart_key_index, current_groq_chart_lock
with current_groq_chart_lock:
if current_groq_chart_key_index >= len(groq_api_keys):
raise Exception("All API keys exhausted for chart generation")
api_key = groq_api_keys[current_groq_chart_key_index]
return ChatGroq(model=model_name, api_key=api_key)
def groq_chart(csv_url: str, question: str):
global current_groq_chart_key_index, current_groq_chart_lock
for attempt in range(len(groq_api_keys)):
try:
# Clean cache before processing
# cache_db_path = "/workspace/cache/cache_db_0.11.db"
# if os.path.exists(cache_db_path):
# try:
# os.remove(cache_db_path)
# except Exception as e:
# logger.info(f"Cache cleanup error: {e}")
data = clean_data(csv_url)
with current_groq_chart_lock:
current_api_key = groq_api_keys[current_groq_chart_key_index]
llm = ChatGroq(model=model_name, api_key=current_api_key)
# Generate unique filename using UUID
chart_filename = f"chart_{uuid.uuid4()}.png"
chart_path = os.path.join("generated_charts", chart_filename)
# Configure SmartDataframe with chart settings
df = SmartDataframe(
data,
config={
'llm': llm,
'save_charts': True, # Enable chart saving
'open_charts': False,
'save_charts_path': os.path.dirname(chart_path), # Directory to save
'custom_chart_filename': chart_filename, # Unique filename
'enable_cache': False
}
)
answer = df.chat(question + instructions)
if process_answer(answer):
return "Chart not generated"
return answer
except Exception as e:
error = str(e)
# if "429" in error:
if error != "":
with current_groq_chart_lock:
current_groq_chart_key_index = (current_groq_chart_key_index + 1)
else:
logger.error(f"Chart generation error: {error}")
return {"error": error}
return {"error": "All API keys exhausted for chart generation"}
# Global locks for key rotation (chart endpoints)
# current_groq_chart_key_index = 0
# current_groq_chart_lock = threading.Lock()
current_langchain_chart_key_index = 0
current_langchain_chart_lock = threading.Lock()
# Use a process pool to run CPU-bound charts generation
process_executor = ProcessPoolExecutor(max_workers=max_cpus-2)
# --- LANGCHAIN-BASED CHART GENERATION ---
def langchain_csv_chart(csv_url: str, question: str, chart_required: bool):
"""
Generate a chart using the langchain-based method.
Modifications:
• No shared deletion of cache.
• Close matplotlib figures after generation.
• Return a list of full chart file paths.
"""
global current_langchain_chart_key_index, current_langchain_chart_lock
data = clean_data(csv_url)
for attempt in range(len(groq_api_keys)):
try:
with current_langchain_chart_lock:
api_key = groq_api_keys[current_langchain_chart_key_index]
current_key = current_langchain_chart_key_index
current_langchain_chart_key_index = (current_langchain_chart_key_index + 1) % len(groq_api_keys)
llm = ChatGroq(model=model_name, api_key=api_key)
tool = PythonAstREPLTool(locals={
"df": data,
"pd": pd,
"np": np,
"plt": plt,
"sns": sns,
"matplotlib": matplotlib,
"uuid": uuid
})
agent = create_pandas_dataframe_agent(
llm,
data,
agent_type="tool-calling",
verbose=True,
allow_dangerous_code=True,
extra_tools=[tool],
return_intermediate_steps=True
)
result = agent.invoke({"input": _prompt_generator(question, True, csv_url)})
output = result.get("output", "")
# Close figures to avoid interference
plt.close('all')
# Extract chart filenames (assuming extract_chart_filenames returns a list)
chart_files = extract_chart_filenames(output)
if len(chart_files) > 0:
# Return full paths (join with your image_file_path)
return [os.path.join(image_file_path, f) for f in chart_files]
if attempt < len(groq_api_keys) - 1:
logger.info(f"Langchain chart error (key {current_key}): {output}")
except Exception as e:
error_message = str(e)
if error_message != "":
with current_langchain_chart_lock:
current_langchain_chart_key_index = (current_langchain_chart_key_index + 1)
logger.warning(f"Rate limit exceeded. Switching to next API key: {groq_api_keys[current_langchain_chart_key_index]}")
else:
logger.error(f"Error with API key {api_key}: {error_message}")
return {"error": error_message}
logger.error("All API keys exhausted for chart generation")
return "Chart generation failed after all retries"
# --- FASTAPI ENDPOINT FOR CHART GENERATION ---
@app.post("/api/csv-chart")
async def csv_chart(request: dict, authorization: str = Header(None)):
"""
Endpoint for generating a chart from CSV data.
This endpoint uses a ProcessPoolExecutor to run the (CPU-bound) chart generation
functions in separate processes so that multiple requests can run in parallel.
"""
# --- Authorization Check ---
if not authorization or not authorization.startswith("Bearer "):
raise HTTPException(status_code=401, detail="Authorization required")
token = authorization.split(" ")[1]
if token != os.getenv("AUTH_TOKEN"):
raise HTTPException(status_code=403, detail="Invalid credentials")
try:
query = request.get("query", "")
csv_url = unquote(request.get("csv_url", ""))
detailed_answer = request.get("detailed_answer", False)
conversation_history = request.get("conversation_history", [])
generate_report = request.get("generate_report", False)
chat_id = request.get("chat_id", "")
if generate_report is True:
report_files = await generate_csv_report(csv_url, query, chat_id, conversation_history)
if report_files is not None:
return {"orchestrator_response": jsonable_encoder(report_files)}
loop = asyncio.get_running_loop()
# First, try the langchain-based method if the question qualifies
if if_initial_chart_question(query):
langchain_result = await loop.run_in_executor(
process_executor, langchain_csv_chart, csv_url, query, True
)
logger.info("Langchain chart result:", langchain_result)
if isinstance(langchain_result, list) and len(langchain_result) > 0:
unique_file_name =f'{str(uuid.uuid4())}.png'
logger.info("Uploading the chart to supabase...")
image_public_url = await upload_file_to_supabase(f"{langchain_result[0]}", unique_file_name, chat_id=chat_id)
logger.info("Image uploaded to Supabase and Image URL is... ", {image_public_url})
os.remove(langchain_result[0])
return {"image_url": image_public_url}
# return FileResponse(langchain_result[0], media_type="image/png")
# Use orchestrator to handle the user's chart query first
if detailed_answer is True:
orchestrator_answer = await asyncio.to_thread(
csv_orchestrator_chat, csv_url, query, conversation_history, chat_id
)
if orchestrator_answer is not None:
return {"orchestrator_response": jsonable_encoder(orchestrator_answer)}
# Next, try the groq-based method
# groq_result = await loop.run_in_executor(
# process_executor, groq_chart, csv_url, query
# )
# logger.info(f"Groq chart result: {groq_result}")
# if isinstance(groq_result, str) and groq_result != "Chart not generated":
# unique_file_name =f'{str(uuid.uuid4())}.png'
# logger.info("Uploading the chart to supabase...")
# image_public_url = await upload_file_to_supabase(f"{groq_result}", unique_file_name, chat_id=chat_id)
# logger.info("Image uploaded to Supabase and Image URL is... ", {image_public_url})
# os.remove(groq_result)
# return {"image_url": image_public_url}
# return FileResponse(groq_result, media_type="image/png")
logger.info("Trying cerebras ai llama...")
result = await query_csv_agent(csv_url, query, chat_id)
logger.info("cerebras ai result ==>", result)
if result is not None and result != "":
return {"orchestrator_response": jsonable_encoder(result)}
# Fallback: try langchain-based again
logger.error("Cerebras ai llama response failed, trying langchain groq....")
langchain_paths = await loop.run_in_executor(
process_executor, langchain_csv_chart, csv_url, query, True
)
logger.info("Fallback langchain chart result:", langchain_paths)
if isinstance(langchain_paths, list) and len(langchain_paths) > 0:
unique_file_name =f'{str(uuid.uuid4())}.png'
logger.info("Uploading the chart to supabase...")
image_public_url = await upload_file_to_supabase(f"{langchain_paths[0]}", unique_file_name, chat_id=chat_id)
logger.info("Image uploaded to Supabase and Image URL is... ", {image_public_url})
os.remove(langchain_paths[0])
return {"image_url": image_public_url}
return FileResponse(langchain_paths[0], media_type="image/png")
else:
logger.error("All chart generation methods failed")
return {"answer": "error"}
except Exception as e:
logger.error(f"Critical chart error: {str(e)}")
return {"answer": "error"}
|