DOU Parser - Extrator do Diário Oficial da União

A client-side web application that parses PDF files from the Brazilian Official Gazette (Diário Oficial da União - DOU) and extracts legal acts into a structured, searchable table.

Features

• Client-side PDF Processing: All parsing happens in the browser using pdfjs-dist
• Smart Document Detection: Automatically identifies different types of legal acts (Portarias, Decretos, etc.)
• Ministry/Organization Tracking: Associates documents with their issuing ministry or organization
• Advanced Filtering: Search across all fields, filter by document type or ministry
• Sortable Table: Click column headers to sort documents
• CSV Export: Export filtered results to CSV format
• Detail Modal: View full document content with a single click
• Responsive Design: Works on desktop, tablet, and mobile devices
• Brazilian Government Aesthetic: Clean, professional UI with government color scheme

How It Works

PDF Upload

1. Drag and drop a DOU PDF file onto the upload zone, or click to select a file
2. The application will process the PDF and extract all text
3. Progress indicator shows processing status

Intelligent Parsing

Sumário Detection

The first pages of DOU PDFs contain a table of contents (Sumário). The parser:

• Detects pages with more than 3 lines containing 10+ consecutive dots
• Automatically skips these summary pages
• Removes any lines with 5+ consecutive dots from content

Document Type Detection

Recognizes the following document types:

• PORTARIA (Nº XXX)
• DECRETO (Nº XXX)
• RESOLUÇÃO (Nº XXX)
• LEI (Nº XXX)
• EDITAL (Nº XXX)
• DESPACHO
• INSTRUÇÃO NORMATIVA (Nº XXX)
• MEDIDA PROVISÓRIA (Nº XXX)
• ATO (Nº XXX)
• ORDEM (Nº XXX)

Ministry/Organization Detection

Automatically identifies and tracks:

• MINISTÉRIO DA/DO [NAME]
• PRESIDÊNCIA DA REPÚBLICA
• CASA CIVIL
• ADVOCACIA-GERAL DA UNIÃO
• CONTROLADORIA-GERAL DA UNIÃO
• AGÊNCIA [NAME]
• INSTITUTO [NAME]
• CONSELHO [NAME]
• SECRETARIA [NAME]
• TRIBUNAL [NAME]
• COMANDO [NAME]

Date Extraction

Extracts dates in Portuguese format: "DD DE [MONTH] DE YYYY"

• Example: "19 DE DEZEMBRO DE 2025" → "19/12/2025"

Signatory Detection

Identifies signatories by:

• Looking for ALL CAPS names
• Checking the last ~10 lines of each document
• Filtering out common headers and footer text
• Requiring reasonable name length (5-100 characters)

Usage

Opening the Application

Simply open index.html in a modern web browser. No server or installation required.

Processing a PDF

1. Upload a DOU PDF file
2. Wait for processing (progress bar will show status)
3. View extracted documents in the table

Filtering Results

• Search: Type in the search box to filter across all fields
• Type Filter: Select a specific document type from the dropdown
• Ministry Filter: Select a specific ministry/organization

Sorting

Click any column header to sort by that column. Click again to reverse sort direction.

Exporting Data

Click the "Exportar CSV" button to download filtered results as a CSV file.

Viewing Full Documents

Click any row in the table to open a modal with the complete document content.

Technical Details

Technology Stack

• HTML5: Semantic markup
• CSS3: Modern styling with flexbox and grid
• Vanilla JavaScript: No framework dependencies
• pdfjs-dist: PDF text extraction library (loaded from CDN)

File Structure

dou-parser/
├── index.html      # Main HTML structure
├── styles.css      # All styling and responsive design
├── app.js          # UI logic, filtering, sorting, export
├── parser.js       # PDF parsing and document extraction
└── README.md       # This file

Browser Compatibility

Works in all modern browsers that support:

• ES6+ JavaScript
• CSS Grid and Flexbox
• Fetch API
• File API

Tested in:

• Chrome 90+
• Firefox 88+
• Safari 14+
• Edge 90+

Color Scheme

The application uses the Brazilian government color scheme:

• Primary (Navy Blue): #003366 - Headers, buttons
• Accent (Yellow): #FFCC00 - Export button, highlights
• Background: #f5f5f5 - Page background
• Text: #333 - Main text color
• White: #ffffff - Cards, table background

Success Criteria Checklist

• ✅ Correctly skips Sumário pages (detects dots pattern)
• ✅ Extracts documents from content pages only
• ✅ Associates correct ministry with each document
• ✅ Clean data without dots artifacts
• ✅ Functional search and filters
• ✅ Working CSV export
• ✅ Professional, clean UI
• ✅ Sortable table columns
• ✅ Document detail modal
• ✅ Responsive design
• ✅ Progress indicator

Example Output

For a document like:

CASA CIVIL

PORTARIAS DE 19 DE DEZEMBRO DE 2025

O MINISTRO DE ESTADO DA CASA CIVIL...

Nº 1.431 - NOMEAR
ALLAN DE ALCÂNTARA, para exercer o cargo...

FLAVIO JOSÉ ROMAN

The parser extracts:

Tipo	Número	Data	Ministério	Página	Prévia	Signatário
PORTARIA	1.431	19/12/2025	CASA CIVIL	2	NOMEAR ALLAN DE ALCÂNTARA...	FLAVIO JOSÉ ROMAN

Development Notes

Parsing Challenges

1. Inconsistent Formatting: DOU PDFs have varying formats across different sections
2. Text Extraction: PDF.js provides basic text extraction; line breaks aren't always reliable
3. Ministry Tracking: Ministries are declared at section headers and apply to subsequent documents
4. Signatory Detection: Heuristic-based approach looking for ALL CAPS names at document end

Future Improvements

• Support for batch PDF processing
• More document types
• Better handling of multi-page documents
• Export to other formats (JSON, Excel)
• Document content search highlighting
• Save/load filtered results
• API integration for automated processing

License

This is a demonstration project created for parsing DOU documents. Use at your own discretion.

Support

For issues or questions, please check the console for error messages and verify that you're using a valid DOU PDF file.