pp-p2p-parser-worker

A Cloudflare Worker for converting an exported P2P account statement .csv into a Portfolio Performance compatible .csv file. The Worker accepts a statement export, detects or uses the selected provider format, applies optional aggregation and returns a downloadable import file for Portfolio Performance.

Table of Contents

• Features
• Quick Start
• Configuration
  • Supported Providers
  • Aggregation Mode
  • Output Language
• How it Works
• Installation & Development
• Testing
• Endpoints
• iFrame Usage
• CLI Usage
• Author & Licence

Features

• Conversion: Upload a .csv export and get a Portfolio Performance compatible .csv file in response
• Multiple Provider and Language Support: Supports multiple P2P lending platforms with language-specific format variants
• Auto-Detection: Detects the likely provider format from the .csv header row when Auto-Detect is selected
• Aggregation Modes: Supports different aggregation modes for bookings of the same type on the same day or month, in addition to exporting each transaction separately
• CLI Compatibility: Keeps a Python local command line parser for existing workflows

Quick Start

You will learn how to deploy the worker to Cloudflare, open the built-in UI, upload a .csv export and download a Portfolio Performance-ready import file.

1. Prepare the Codebase

Cloudflare always requires a code source (repository or ZIP) to deploy a Worker. Choose one of the following:

• Git (recommended): Fork this repository into your own GitHub/GitLab account.
• ZIP (manual upload): Download the code as a ZIP file and prepare it for upload.

2. Add a Worker in Cloudflare

1. Log in to the Cloudflare Dashboard.
2. Navigate to Workers & Pages -> Workers and create a new Worker with the name pp-p2p-parser-worker.

3. Deploy

• Cloudflare Git integration (recommended): Connect your forked repository directly to your GitHub/GitLab account in the Cloudflare Dashboard. Cloudflare will build and deploy automatically. Additional information about the Git integration for Cloudflare Workers can be found in the Cloudflare documentation.
• ZIP (manual upload): Upload your prepared ZIP file using the Dashboard’s editor or deployment UI.
• Wrangler: Deploy from your local checkout with npm run deploy.
• CI pipeline: If you want to have control over the CI checks and deployment, you can set up your own pipeline that runs the tests and deploys to Cloudflare. You can use the same CI configuration as this repository, which is available in .github/workflows/ci.yml. Make sure to update the deployment step with your own Cloudflare API credentials and Worker name. Additional information on setting up CI/CD pipelines for Cloudflare Workers can be found in the Cloudflare documentation.

Important: When using Cloudflare Git integration, go to Settings -> Build -> Branch Control in your Worker project. Make sure to deactivate (uncheck) the option for enabling builds for non-production branches. If this setting is active, any push to your develop (or other non-production) branch will trigger a deployment to your Worker, which may not be desired for production stability.

4. Open the UI and use the Service

Open:

https://<your-worker>.<your-subdomain>.workers.dev/

Upload a .csv export from one of the supported providers, select the provider or keep Auto-Detect, choose the Portfolio Performance language and aggregation mode, then click Convert to download a compatible .csv file.

Configuration

Supported Providers

Provider Key	Format
bondora	Bondora account statement
bondora_go_grow	Bondora Go & Grow account statement
debitumnetwork	Debitum Network account statement
estateguru_de	EstateGuru German export
estateguru_de_legacy	EstateGuru older German export
estateguru_en	EstateGuru English export
lande	Lande account statement
mintos_de	Mintos German export
mintos_en	Mintos English export
robocash	Robocash account statement
swaper	Swaper account statement
viainvest	Viainvest account statement

Aggregation Mode

Mode	Behavior	Best Use
transaction	Exports each supported booking as a separate row.	Default and recommended mode when you want a detailed import history.
daily	Groups bookings with the same date, transaction type and currency into one row per day.	Reducing many small transactions while keeping daily totals.
monthly	Groups bookings with the same transaction type and currency into one row per month, dated on the last day of the month.	Compact long-term imports. Be careful with the current month because Portfolio Performance may ignore future-dated rows.

Monthly aggregation can create booking dates later than the import date for the current month. Portfolio Performance may ignore future-dated rows.

Output Language

• de: German column names and transaction types. This is the default for backwards compatibility.
• en: English column names and transaction types.

How it Works

Provider Detection

P2P platforms export account statements in different CSV formats. Column names, date formats, transaction descriptions and language variants differ between providers. When the user selects a file, the browser reads the CSV header row and tries to detect the provider locally by comparing the uploaded headers with the required columns from the supported provider configurations.

If detection succeeds, the provider dropdown is updated automatically. If detection fails, the UI keeps Auto-Detect selected and shows a warning so the user can choose a provider manually.

If a provider is selected manually, the Worker verifies that the required columns exist before parsing. If the headers look like another provider, the error message names the likely match.

Aggregation

This Worker supports three aggregation modes:

Mode	Behavior
transaction	Each supported input row is converted directly into one Portfolio Performance output row. Keeps the original booking date, value, currency, mapped transaction type and generated note.
daily	Supported rows are grouped by booking date and mapped transaction type. Values in each group are summed. The output note is Tageszusammenfassung in German output and Daily summary in English output. The first currency seen for the group is used.
monthly	Supported rows are grouped by month and mapped transaction type. Values in each group are summed and the output date is set to the last day of that month. The output note is Monatszusammenfassung in German output and Monthly summary in English output. Monthly rows for the current month may be dated in the future and can be ignored by Portfolio Performance.

Portfolio Performance Language

The parser uses an English internal transaction model: Date, Value, Currency, Type and Note, with transaction types such as Deposit, Withdrawal, Interest and Fees. At the final CSV writing step, the selected output language controls the exported column headers, transaction type labels and generated aggregation notes. Provider detection and parsing are independent of the output language.

Output Language	Header	Transaction Types
de	Datum,Wert,Buchungswährung,Typ,Notiz	Einlage, Entnahme, Zinsen, Gebühren
en	Date,Value,Transaction Currency,Type,Note	Deposit, Withdrawal, Interest, Fees

Conversion

When the user clicks Convert, the browser sends the CSV file and selected options to POST /parse. The Worker validates the provider, aggregation mode, output language and CSV content.

Each supported row is mapped to a Portfolio Performance account transaction category by matching the provider-specific transaction description against configured regular expressions. Unsupported or intentionally ignored transaction descriptions are skipped. Input values with comma or dot decimal separators, including common thousands-separator formats, are normalized before conversion. The provider currency column is used when available. Providers without a currency column default to EUR.

Finally, the Worker returns a downloadable Portfolio Performance CSV file. If an error occurs, it is returned as JSON and shown in the UI as a red status message.

Installation & Development

1. Clone the Repository

   git clone https://github.com/jasonhaak/pp-p2p-parser-worker.git
   cd pp-p2p-parser-worker

2. Install Dependencies There are no project-local npm dependencies and no required Python runtime dependencies for the Worker. The Worker uses Cloudflare's built-in Python runtime SDK via the disable_python_external_sdk compatibility flag in wrangler.toml.

   For local development you need:

   • Node.js with npx to run Wrangler
   • Python 3 to run tests and local parser commands

3. Run the Worker Locally

   npx wrangler dev

   Open the local UI at:

   http://localhost:8787
   

Provider Configuration

The provider configuration files are located in config. Each provider has a .yml file defining the expected .csv format, column mappings and parsing rules. You can add new providers by creating a new configuration file following the existing examples.

The .yml provider configuration is converted in the CI pipeline with the help of tools/generate_provider_configs.py into the src/provider_configs.py module. This module is imported by the Worker to access the provider definitions for parsing and conversion.

Testing

This project uses Python's built-in unittest for unit tests. Test data for each provider is located in test/testdata. Additionally, the CI pipeline runs syntax checks for converting and validating the configuration files for each provider, Worker bundle validation, Python and JavaScript code as well as code coverage.

Run the suite locally to test core parsing and conversion logic:

python3 -m unittest discover -s test

Check Python syntax for all source files:

python3 -m py_compile src/*.py tools/*.py

Check frontend JavaScript syntax for testing the UI:

node --check public/app.js

Endpoints

Route	Description
GET /	Serves the static upload UI
POST /parse	Accepts a .csv export and returns a Portfolio Performance compatible .csv file in response

/parse Query Parameters

Parameter	Required	Default	Description
file	Yes	N/A	The .csv file to be converted, sent as multipart form data
provider	No	auto	The provider format to use for parsing. If auto is selected, the Worker will attempt to detect the provider format from the header row of the uploaded .csv file.
aggregate	No	transaction	The aggregation mode to apply to the output file.
output_language	No	de	The language to use for column names and transaction types in the output file.

Example

Example curl request for the API endpoint:

curl -X POST http://localhost:8787/parse \
  -F "file=@test/testdata/mintos.csv" \
  -F "provider=auto" \
  -F "aggregate=transaction" \
  -F "output_language=en" \
  -o /tmp/portfolio_performance.csv

Successful responses return text/csv with a Content-Disposition download filename.

iFrame Usage

The UI is designed to be embedded in another page. It automatically switches to embedded mode when loaded inside an iframe. You can also force embedded mode with /embed or /?embed=1.

Recommended iFrame embedding:

iframe {
  width: 100%;
  min-height: 720px;
  border: 0;
  background: #F6F6F6;
  display: block;
}

The embed view removes the footer, outer spacing, shadow and rounded outer frame.

The UI response also sends:

Content-Security-Policy: frame-ancestors *

Adjust this header before production use if you only want specific domains to embed the UI.

CLI Usage

The local CLI remains available. It accepts the same provider, aggregation and output language options as the Worker. The CLI writes the converted file next to the input file with the name portfolio_performance__<provider>__<language>.csv:

python3 -m src.cli --type mintos_en test/testdata/mintos.csv

Available aggregation modes:

python3 -m src.cli --type mintos_en --aggregate daily test/testdata/mintos.csv

Available output languages:

python3 -m src.cli --type mintos_en --output-language en test/testdata/mintos.csv

Author & Licence

This project is maintained by Jason Haak and is licensed under the GPL-3.0 licence.

The original parser work comes from ChrisRBe's PP-P2P-Parser repository. This project adapts that parser into a Cloudflare Worker.

This project is not affiliated with Portfolio Performance or any P2P lending platform.