ahmed-bakr1988/SaaS-PDF
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Unify task status error schema and classify PDF AI failures

Browse Source
This commit is contained in:
Ahmed Bakr Ali
2026-03-24 23:30:46 +02:00
parent a6e0cab0b3
commit b2a7678848
5 changed files with 232 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files

85
backend/app/services/pdf_ai_service.py
View File

@@ -11,7 +11,16 @@ logger = logging.getLogger(__name__)
class PdfAiError(Exception):
"""Custom exception for PDF AI service failures."""
pass
def __init__(
self,
user_message: str,
error_code: str = "PDF_AI_ERROR",
detail: str | None = None,
):
super().__init__(user_message)
self.user_message = user_message
self.error_code = error_code
self.detail = detail
def _estimate_tokens(text: str) -> int:
@@ -33,7 +42,11 @@ def _extract_text_from_pdf(input_path: str, max_pages: int = 50) -> str:
texts.append(f"[Page {i + 1}]\n{text}")
return "\n\n".join(texts)
except Exception as e:
raise PdfAiError(f"Failed to extract text from PDF: {str(e)}")
raise PdfAiError(
"Failed to extract text from PDF.",
error_code="PDF_TEXT_EXTRACTION_FAILED",
detail=str(e),
)
def _call_openrouter(
@@ -48,7 +61,10 @@ def _call_openrouter(
from app.services.ai_cost_service import check_ai_budget, AiBudgetExceededError
check_ai_budget()
except AiBudgetExceededError:
raise PdfAiError("Monthly AI processing budget has been reached. Please try again next month.")
raise PdfAiError(
"Monthly AI processing budget has been reached. Please try again next month.",
error_code="AI_BUDGET_EXCEEDED",
)
except Exception:
pass # Don't block if cost service unavailable
@@ -57,7 +73,8 @@ def _call_openrouter(
if not settings.api_key:
logger.error("OPENROUTER_API_KEY is not set or is a placeholder value.")
raise PdfAiError(
"AI features are temporarily unavailable. Our team has been notified."
"AI features are temporarily unavailable. Our team has been notified.",
error_code="OPENROUTER_MISSING_API_KEY",
)
messages = [
@@ -84,25 +101,29 @@ def _call_openrouter(
if response.status_code == 401:
logger.error("OpenRouter API key is invalid or expired (401).")
raise PdfAiError(
"AI features are temporarily unavailable due to a configuration issue. Our team has been notified."
"AI features are temporarily unavailable due to a configuration issue. Our team has been notified.",
error_code="OPENROUTER_UNAUTHORIZED",
)
if response.status_code == 402:
logger.error("OpenRouter account has insufficient credits (402).")
raise PdfAiError(
"AI processing credits have been exhausted. Please try again later."
"AI processing credits have been exhausted. Please try again later.",
error_code="OPENROUTER_INSUFFICIENT_CREDITS",
)
if response.status_code == 429:
logger.warning("OpenRouter rate limit reached (429).")
raise PdfAiError(
"AI service is experiencing high demand. Please wait a moment and try again."
"AI service is experiencing high demand. Please wait a moment and try again.",
error_code="OPENROUTER_RATE_LIMIT",
)
if response.status_code >= 500:
logger.error("OpenRouter server error (%s).", response.status_code)
raise PdfAiError(
"AI service provider is experiencing issues. Please try again shortly."
"AI service provider is experiencing issues. Please try again shortly.",
error_code="OPENROUTER_SERVER_ERROR",
)
response.raise_for_status()
@@ -112,7 +133,11 @@ def _call_openrouter(
if data.get("error"):
error_msg = data["error"].get("message", "") if isinstance(data["error"], dict) else str(data["error"])
logger.error("OpenRouter returned an error payload: %s", error_msg)
raise PdfAiError("AI service encountered an issue. Please try again.")
raise PdfAiError(
"AI service encountered an issue. Please try again.",
error_code="OPENROUTER_ERROR_PAYLOAD",
detail=error_msg,
)
reply = (
data.get("choices", [{}])[0]
@@ -122,7 +147,10 @@ def _call_openrouter(
)
if not reply:
raise PdfAiError("AI returned an empty response. Please try again.")
raise PdfAiError(
"AI returned an empty response. Please try again.",
error_code="OPENROUTER_EMPTY_RESPONSE",
)
# Log usage
try:
@@ -142,13 +170,23 @@ def _call_openrouter(
except PdfAiError:
raise
except requests.exceptions.Timeout:
raise PdfAiError("AI service timed out. Please try again.")
raise PdfAiError(
"AI service timed out. Please try again.",
error_code="OPENROUTER_TIMEOUT",
)
except requests.exceptions.ConnectionError:
logger.error("Cannot connect to OpenRouter API at %s", settings.base_url)
raise PdfAiError("AI service is unreachable. Please try again shortly.")
raise PdfAiError(
"AI service is unreachable. Please try again shortly.",
error_code="OPENROUTER_CONNECTION_ERROR",
)
except requests.exceptions.RequestException as e:
logger.error("OpenRouter API error: %s", e)
raise PdfAiError("AI service is temporarily unavailable.")
raise PdfAiError(
"AI service is temporarily unavailable.",
error_code="OPENROUTER_REQUEST_ERROR",
detail=str(e),
)
# ---------------------------------------------------------------------------
@@ -166,11 +204,11 @@ def chat_with_pdf(input_path: str, question: str) -> dict:
{"reply": "...", "pages_analyzed": int}
"""
if not question or not question.strip():
raise PdfAiError("Please provide a question.")
raise PdfAiError("Please provide a question.", error_code="PDF_AI_INVALID_INPUT")
text = _extract_text_from_pdf(input_path)
if not text.strip():
raise PdfAiError("Could not extract any text from the PDF.")
raise PdfAiError("Could not extract any text from the PDF.", error_code="PDF_TEXT_EMPTY")
# Truncate to fit context window
max_chars = 12000
@@ -206,7 +244,7 @@ def summarize_pdf(input_path: str, length: str = "medium") -> dict:
"""
text = _extract_text_from_pdf(input_path)
if not text.strip():
raise PdfAiError("Could not extract any text from the PDF.")
raise PdfAiError("Could not extract any text from the PDF.", error_code="PDF_TEXT_EMPTY")
length_instruction = {
"short": "Provide a brief summary in 2-3 sentences.",
@@ -245,11 +283,11 @@ def translate_pdf(input_path: str, target_language: str) -> dict:
{"translation": "...", "pages_analyzed": int, "target_language": str}
"""
if not target_language or not target_language.strip():
raise PdfAiError("Please specify a target language.")
raise PdfAiError("Please specify a target language.", error_code="PDF_AI_INVALID_INPUT")
text = _extract_text_from_pdf(input_path)
if not text.strip():
raise PdfAiError("Could not extract any text from the PDF.")
raise PdfAiError("Could not extract any text from the PDF.", error_code="PDF_TEXT_EMPTY")
max_chars = 10000
truncated = text[:max_chars]
@@ -325,7 +363,8 @@ def extract_tables(input_path: str) -> dict:
if not result_tables:
raise PdfAiError(
"No tables found in the PDF. This tool works best with PDFs containing tabular data."
"No tables found in the PDF. This tool works best with PDFs containing tabular data.",
error_code="PDF_TABLES_NOT_FOUND",
)
logger.info(f"Extracted {len(result_tables)} tables from PDF")
@@ -338,6 +377,10 @@ def extract_tables(input_path: str) -> dict:
except PdfAiError:
raise
except ImportError:
raise PdfAiError("tabula-py library is not installed.")
raise PdfAiError("tabula-py library is not installed.", error_code="TABULA_NOT_INSTALLED")
except Exception as e:
raise PdfAiError(f"Failed to extract tables: {str(e)}")
raise PdfAiError(
"Failed to extract tables.",
error_code="PDF_TABLE_EXTRACTION_FAILED",
detail=str(e),
)