Skip to content

fix: improvments for docs and warning #100

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Abdeali099 merged 4 commits into from
Apr 16, 2026
Merged

fix: improvments for docs and warning #100

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
0e54fc3
fix: update README to include Anthropic model and correct model names
Abdeali099 Apr 16, 2026
4a82d26
fix: handle missing docling dependency with informative error message
Abdeali099 Apr 16, 2026
356e213
fix: enhance setup instructions for Docling, PDFtoText, and OCRMyPDF …
Abdeali099 Apr 16, 2026
4ebb383
fix: add warning for PDF processor changes with setup instructions
Abdeali099 Apr 16, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
39 changes: 22 additions & 17 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@ Transaction Parser is an AI-powered add-on for ERPNext that automatically extrac

## Features

**AI-Powered Extraction**: Uses advanced AI models (OpenAI, DeepSeek, Google Gemini) to extract structured data from PDFs
**AI-Powered Extraction**: Uses advanced AI models (OpenAI, DeepSeek, Google Gemini, Anthropic) to extract structured data from PDFs
* **Multi-Document Support**: Handles Sales Orders and Purchase Invoices (Expenses)
* **Regional Support**: Special handling for India-specific requirements (GSTIN, PAN, HSN codes)
* **Email Integration**: Automatically processes documents from incoming emails
Expand All @@ -29,20 +29,23 @@ Navigate to **Transaction Parser Settings** and configure:
* OpenAI gpt-4o
* OpenAI gpt-4o-mini
* OpenAI gpt-5
* Google Gemini 2.5 pro
* Google Gemini 2.5 flash
* OpenAI gpt-5-mini
* Google Gemini Pro-2.5
* Google Gemini Flash-2.5
* Claude Haiku-4.5

<img width="773" height="291" alt="image" src="https://github.com/user-attachments/assets/fc40bea1-1e11-4ef3-bcdf-f6c1db8585c8" />

3\. API Keys Setup

Add your API keys for the AI services:

| Service Provider | Models Supported |
|:-----------------|:---------------------------------|
| OpenAI | gpt-4o, gpt-4o-mini , gpt-5 |
| DeepSeek | deepseek-chat, deepseek-reasoner |
| Google | gemini 2.5 pro, gemini 2.5 flash |
| Service Provider | Models Supported |
|:-----------------|:---------------------------------------|
| OpenAI | gpt-4o, gpt-4o-mini, gpt-5, gpt-5-mini |
| DeepSeek | deepseek-chat, deepseek-reasoner |
| Google | gemini-2.5-pro, gemini-2.5-flash |
| Anthropic | claude-haiku-4-5 |

<img width="800" height="148" alt="image" src="https://github.com/user-attachments/assets/77f30bd8-59a1-4b66-8bf4-964bc2347ce4" />

Expand Down Expand Up @@ -94,15 +97,17 @@ When enabled, the system automatically:

## Model Comparison

| Model | Provider | Best For | Speed | Cost |
|:------------------|:---------|:--------------------------------------------------------|:----------|:------------|
| gpt-5 | OpenAI | State-of-the-art accuracy, complex multi-page documents | Medium | High |
| gpt-4o | OpenAI | Complex documents, high accuracy | Medium | Medium-High |
| gpt-4o-mini | OpenAI | Cost-effective, good accuracy | Fast | Low |
| gemini-2.5-pro | Google | Advanced reasoning, large context window | Medium | Medium |
| gemini-2.5-flash | Google | Fast processing, bulk documents | Very Fast | Low |
| deepseek-chat | DeepSeek | General purpose extraction | Fast | Low |
| deepseek-reasoner | DeepSeek | Complex reasoning tasks | Slow | Medium |
| Model | Provider | Best For | Speed | Cost |
|:------------------|:----------|:--------------------------------------------------------|:----------|:------------|
| gpt-5 | OpenAI | State-of-the-art accuracy, complex multi-page documents | Medium | High |
| gpt-5-mini | OpenAI | Efficient reasoning, cost-effective | Fast | Medium |
| gpt-4o | OpenAI | Complex documents, high accuracy | Medium | Medium-High |
| gpt-4o-mini | OpenAI | Cost-effective, good accuracy | Fast | Low |
| gemini-2.5-pro | Google | Advanced reasoning, large context window | Medium | Medium |
| gemini-2.5-flash | Google | Fast processing, bulk documents | Very Fast | Low |
| deepseek-chat | DeepSeek | General purpose extraction | Fast | Low |
| deepseek-reasoner | DeepSeek | Complex reasoning tasks | Slow | Medium |
| claude-haiku-4-5 | Anthropic | Fast, lightweight tasks | Fast | Low |

## India-Specific Features

Expand Down
17 changes: 17 additions & 0 deletions ...ser/transaction_parser/doctype/transaction_parser_settings/transaction_parser_settings.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,6 +63,23 @@ def validate(self):
self.validate_incoming_email_accounts()
self.validate_party_email()
self.validate_json_fields()
self.warn_on_pdf_processor_change()

def warn_on_pdf_processor_change(self):
if not self.has_value_changed("pdf_processor"):
return

frappe.msgprint(
_(
"Make sure the required dependencies for {0} are installed.<br>"
"See {1} for setup instructions."
).format(
frappe.bold(self.pdf_processor),
'<a href="https://github.com/resilient-tech/transaction-parser#pdf-processor-setup" target="_blank">PDF Processor Setup</a>',
),
title=_("PDF Processor Changed"),
indicator="orange",
)

def validate_lookback_count(self):
if self.invoice_lookback_count <= 0:
Expand Down
54 changes: 40 additions & 14 deletions transaction_parser/transaction_parser/utils/pdf_processor.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -95,7 +95,10 @@ class DoclingPDFProcessor(PDFProcessor):

_converter = None

# TODO: Give detail of install `docling` system dependency and opencv-python-headless for OCR
SETUP_URL = (
"https://github.com/resilient-tech/transaction-parser#3-docling-optional"
)

def process(self, file: io.BytesIO | File, page_limit: int | None = None) -> str:
try:
from docling.datamodel.base_models import ConversionStatus, DocumentStream
Expand All @@ -104,8 +107,9 @@ def process(self, file: io.BytesIO | File, page_limit: int | None = None) -> str
title=_("Missing Dependency"),
msg=_(
"docling is not installed.<br>"
"Install it with: <code>bench pip install transaction_parser[docling]</code>"
),
"Install it with: <code>bench pip install transaction_parser[docling]</code><br>"
"See <a href='{0}'>setup instructions</a> for more details."
).format(self.SETUP_URL),
)

file = self.get_sanitized_file(file, page_limit)
Expand All @@ -132,12 +136,25 @@ def process(self, file: io.BytesIO | File, page_limit: int | None = None) -> str

def _get_converter(self):
if DoclingPDFProcessor._converter is None:
from docling.datamodel.base_models import InputFormat
from docling.datamodel.pipeline_options import (
EasyOcrOptions,
PdfPipelineOptions,
)
from docling.document_converter import DocumentConverter, PdfFormatOption
try:
from docling.datamodel.base_models import InputFormat
from docling.datamodel.pipeline_options import (
EasyOcrOptions,
PdfPipelineOptions,
)
from docling.document_converter import (
DocumentConverter,
PdfFormatOption,
)
except ImportError:
frappe.throw(
title=_("Missing Dependency"),
msg=_(
"docling is not installed.<br>"
"Install it with: <code>bench pip install transaction_parser[docling]</code><br>"
"See <a href='{0}'>setup instructions</a> for more details."
).format(self.SETUP_URL),
)
Comment thread
Abdeali099 marked this conversation as resolved.

pipeline_options = PdfPipelineOptions()
pipeline_options.do_ocr = True
Expand All @@ -157,6 +174,10 @@ class PDFtoTextProcessor(PDFProcessor):
PDF processor using pdftotext for layout-preserving text extraction.
"""

SETUP_URL = (
"https://github.com/resilient-tech/transaction-parser#1-pdftotext-default"
)

def process(self, file: io.BytesIO | File, page_limit: int | None = None) -> str:
file = self.get_sanitized_file(file, page_limit)
return self.get_text(file)
Expand All @@ -171,8 +192,9 @@ def get_text(self, file: io.BytesIO) -> str:
"pdftotext is not installed.<br>"
"Install OS dependencies first if not already installed: "
"<code>sudo apt install build-essential libpoppler-cpp-dev pkg-config python3-dev</code>"
"<br>Then run: <code>bench setup requirements</code>"
),
"<br>Then run: <code>bench setup requirements</code><br>"
"See <a href='{0}'>setup instructions</a> for more details."
).format(self.SETUP_URL),
)

pdf = pdftotext.PDF(file, physical=True)
Expand All @@ -185,13 +207,16 @@ class OCRMyPDFProcessor(PDFProcessor):
PDF processor using PyMuPDF for text extraction and OCRMyPDF for OCR.
"""

SETUP_URL = (
"https://github.com/resilient-tech/transaction-parser#2-ocrmypdf-optional"
)

def process(self, file: io.BytesIO | File, page_limit: int | None = None) -> str:
file = self.get_sanitized_file(file, page_limit)
file = self.apply_ocr(file)

return self.get_text(file)

# TODO: Give detail of install `tesseract-ocr` system dependency
def apply_ocr(self, file: io.BytesIO) -> io.BytesIO:
try:
import ocrmypdf
Expand All @@ -200,8 +225,9 @@ def apply_ocr(self, file: io.BytesIO) -> io.BytesIO:
title=_("Missing Dependency"),
msg=_(
"ocrmypdf is not installed.<br>"
"Install it with: <code>bench pip install transaction_parser[ocrmypdf]</code>"
),
"Install it with: <code>bench pip install transaction_parser[ocrmypdf]</code><br>"
"See <a href='{0}'>setup instructions</a> for more details."
).format(self.SETUP_URL),
)

file.seek(0)
Expand Down
Loading