File size: 14,779 Bytes
4fbcf68 a8dba50 4fbcf68 a8dba50 4fbcf68 a8dba50 4fbcf68 a8dba50 4fbcf68 a8dba50 4fbcf68 a8dba50 4fbcf68 a8dba50 4fbcf68 a8dba50 4fbcf68 a8dba50 4fbcf68 a8dba50 4fbcf68 a8dba50 4fbcf68 a8dba50 |
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 |
# Import necessary modules
import asyncio
import os
import threading
import uuid
from fastapi.encoders import jsonable_encoder
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
from csv_service import clean_data, extract_chart_filenames
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 supabase_service import upload_image_to_supabase
from util_service import _prompt_generator, process_answer
import matplotlib
matplotlib.use('Agg')
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(",")
# 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
class CsvCommonHeadersRequest(BaseModel):
file_urls: list[str]
class CsvsMergeRequest(BaseModel):
file_urls: list[str]
merge_type: str
common_columns_name: list[str]
# 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()
# CHAT CODING STARTS FROM HERE
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
# 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:
# Delete cache file if exists
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:
print(f"Error deleting cache DB file: {e}")
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
}
)
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 "429" in error_message:
with current_groq_key_lock:
current_groq_key_index += 1
if current_groq_key_index >= len(groq_api_keys):
print("All API keys exhausted.")
return None
else:
print(f"Error with API key index {current_groq_key_index}: {error_message}")
return None
# 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="openai-tools",
verbose=True,
allow_dangerous_code=True,
extra_tools=[tool],
return_intermediate_steps=True
)
prompt = _prompt_generator(question, chart_required)
result = agent.invoke({"input": prompt})
return result.get("output")
except Exception as e:
print(f"Error with key index {current_key}: {str(e)}")
# If all keys are exhausted, return None
print("All API keys have been exhausted.")
return None
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, arrange them in a grid format (2x2, 3x3, etc.)
- 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:
print(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
}
)
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:
with current_groq_chart_lock:
current_groq_chart_key_index = (current_groq_chart_key_index + 1) % len(groq_api_keys)
else:
print(f"Chart generation error: {error}")
return {"error": error}
print("All API keys exhausted for chart generation")
return None
def langchain_csv_chart(csv_url: str, question: str, chart_required: bool):
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="openai-tools",
verbose=True,
allow_dangerous_code=True,
extra_tools=[tool],
return_intermediate_steps=True
)
result = agent.invoke({"input": _prompt_generator(question, True)})
output = result.get("output", "")
# Verify chart file creation
chart_files = extract_chart_filenames(output)
if len(chart_files) > 0:
return chart_files
if attempt < len(groq_api_keys) - 1:
print(f"Langchain chart error (key {current_key}): {output}")
except Exception as e:
print(f"Langchain chart error (key {current_key}): {str(e)}")
print("All API keys exhausted for chart generation")
return None
###########################################################################################################################
async def csv_chart(csv_url: str, query: str):
try:
# First try Groq-based chart generation
try:
groq_result = await asyncio.to_thread(groq_chart, csv_url, query)
print(f"Generated Chart (Groq): {groq_result}")
if groq_result != 'Chart not generated':
unique_file_name = f'{str(uuid.uuid4())}.png'
image_public_url = await upload_image_to_supabase(groq_result, unique_file_name)
print(f"Image uploaded to Supabase: {image_public_url}")
return {"image_url": image_public_url}
except Exception as groq_error:
print(f"Groq chart generation failed, falling back to Langchain: {str(groq_error)}")
# Fallback to Langchain if Groq fails
try:
langchain_paths = await asyncio.to_thread(langchain_csv_chart, csv_url, query, True)
print("Fallback langchain chart result:", langchain_paths)
if isinstance(langchain_paths, list) and len(langchain_paths) > 0:
unique_file_name = f'{str(uuid.uuid4())}.png'
print("Uploading the chart to supabase...")
image_public_url = await upload_image_to_supabase(langchain_paths[0], unique_file_name)
print("Image uploaded to Supabase and Image URL is... ", image_public_url)
return {"image_url": image_public_url}
except Exception as langchain_error:
print(f"Langchain chart generation also failed: {str(langchain_error)}")
# If both methods fail
return {"error": "All chart generation methods failed"}
except Exception as e:
print(f"Critical chart error: {str(e)}")
return {"error": "Internal system error"}
async def csv_chat(csv_url: str, query: str):
try:
updated_query = f"{query} and Do not show any charts or graphs."
# Process with langchain_chat first
try:
lang_answer = await asyncio.to_thread(
langchain_csv_chat, csv_url, query, False
)
if not process_answer(lang_answer):
return {"answer": jsonable_encoder(lang_answer)}
return {"answer": "Sorry, I couldn't find relevant data..."}
except Exception as langchain_error:
print(f"Langchain error, falling back to Groq: {str(langchain_error)}")
# Process with groq_chat if langchain fails
try:
groq_answer = await asyncio.to_thread(groq_chat, csv_url, updated_query)
print("groq_answer:", groq_answer)
if process_answer(groq_answer) == "Empty response received.":
return {"answer": "Sorry, I couldn't find relevant data..."}
if process_answer(groq_answer):
return {"answer": "error"}
return {"answer": jsonable_encoder(groq_answer)}
except Exception as groq_error:
print(f"Groq processing error: {str(groq_error)}")
return {"answer": "error"}
except Exception as e:
print(f"Error processing request: {str(e)}")
return {"answer": "error"} |