🗂️ Excel → A4 PDF Print Studio

Turn messy government Excel sheets into beautiful, print-ready A4 PDFs — correct Devanagari, hidden columns dropped, one click. No Excel gymnastics, no re-typing.

---

✨ What makes this brilliant?

Every district office has the same spreadsheet problem: a 1000-row Excel workbook, thirteen visible columns, five hidden legacy columns, a legacy KrutiDev-encoded Hindi font, and a signature block that must start on a fresh page. Printing it means half a day of column juggling and still-mangled Devanagari.

Excel → A4 PDF Print Studio reads the workbook, profiles each sheet (titles, header band, signature lines, numeric density, KrutiDev ratio), suggests the right layout, and renders a clean A4 PDF — one page per signature block, with every KrutiDev byte converted to real Unicode Devanagari. You can tune each sheet in a buttoned browser UI and save the recipe to layouts.json. It also still runs as a plain one-command CLI. 🇮🇳📄

Built for Public Works Department (PWD) budget estimates — sane for any awkward workbook.

---

🚀 Key Features

• 🧠 Automatic sheet profiling — detects title rows, the column-header band, signature lines and numeric density with zero configuration.
• 🕵️ Live hidden-column detection — hidden columns (M N O P Q, or anything else) are found per sheet at runtime, never hard-coded.
• 🔤 KrutiDev → Unicode converter — legacy Remington/Walkman-Chanakya Hindi text is decoded to real Devanagari, so headings render instead of lM+d gibberish.
• 🅰️ Smart English vs. KrutiDev classifier — genuine English words, work-IDs and numbers are left alone, never mangled.
• 🧾 Six layout presets — cover letter, budget details, continuous flow table, summary, plain grid, and auto-detect.
• 🖨️ Pixel-precise pagination — ReportLab LongTable with a correctly-bounded repeatRows, so 70-page tables split cleanly instead of throwing LayoutError.
• 🎛️ Buttoned web UI — one card per worksheet, per-sheet config drawer, live page-1 preview thumbnails.
• 💾 Reproducible configs — every choice round-trips through layouts.json (save, reload, re-render).
• 🧩 Non-destructive — uploads are copied to a scratch folder; your INPUT/ folder is never touched.
• 🖱️ Windows-friendly launchers — run_pdf.bat auto-installs dependencies and opens the output folder.

---

📸 Demo / Screenshots

Slots ready — drop your images in docs/screenshots/ and they will appear here. Screenshot shots render in the code below (currently commented out so nothing shows broken).

Suggested captures: the drop-zone (1-upload.png), the per-sheet board with badges (2-sheet-board.png), the config drawer (3-config-drawer.png), and a rendered Hindi budget page (4-preview.png).

💡 Want maximum reach? Record a 20–40 s screen capture of “drop a workbook → Render All → open the PDF” and embed it as a GIF at the top of this README.

---

⚡ Quick Start (under 90 seconds)

Prerequisites: Python 3.9+ and pip.

# 1 — clone
git clone https://github.com/CRAJKUMARSINGH/PDFExcelExtract.git
cd PDFExcelExtract

# 2 — install dependencies
python -m pip install -r requirements.txt

# 3 — run the studio (web UI)
python app.py
# → open http://127.0.0.1:8000

Prefer the command line? Render a workbook straight to PDF:

python apply.py -i "INPUT/your-workbook.xlsx" -o "OUTPUT/budget_paged.pdf"

Windows, zero terminal: double-click run_pdf.bat — it checks Python, installs the packages, renders, and opens OUTPUT\.

---

📖 How to Use

Option A — the buttoned web UI (recommended)

Step	What happens
1. Upload	Drop an .xlsx / .xls into the drop-zone (or click Choose Excel).
2. Sheet Board	One card per worksheet with badges: rows×cols, titles, signatures, hidden M N O P Q, kruti 62% — plus the auto-detected layout.
3. Configure	Open the drawer and tune Layout · Columns · Rows & Breaks · Language & Font for that one sheet.
4. Preview / Render	Preview page 1 for a thumbnail, Render this sheet for one PDF, or Render All for the whole workbook.
5. Save the recipe	💾 Save layouts.json persists every choice for next time.

Option B — the CLI

python apply.py --report                              # dry-run: list every page block
python apply.py -s Details                            # render one named sheet
python apply.py --all-sheets                          # render EVERY sheet -> OUTPUT/<sheet>_paged.pdf
python apply.py --keep-hidden                         # include hidden columns M N O P Q
python apply.py -i INPUT/wb.xls -o OUTPUT/out.pdf     # explicit input / output
python apply.py --margin 10 --font-dir fonts          # 10 mm margins, custom font folder

--report is the fastest way to sanity-check a new workbook: it prints the detected page blocks without writing a PDF.

---

🔧 Configuration / Options

CLI flags (apply.py)

Flag	Default	Description
-i, --input	INPUT/FINAL Budget Anuman 31 October 2026.xlsx	Input workbook (.xls or .xlsx)
-o, --output	OUTPUT/budget_paged.pdf	Output PDF path
-s, --sheet	Details	Worksheet to read
--all-sheets	off	Render every worksheet to OUTPUT/<sheet>_paged.pdf
--keep-hidden	off	Include the hidden columns (M–Q)
--margin MM	12	Page margin in millimetres
--font-dir DIR	fonts	Folder holding NotoSansDevanagari-Regular.ttf
--report	off	Dry-run: print the page-block list, write nothing

Layout presets (engine/layouts.py)

Preset id	English / हिंदी	Used for
budget-details	Budget details / बजट विवरण	One page per signature block, ruled title band
flow-table	Continuous table / सतत सारणी	Long sheet, header repeated every page
cover-letter	Cover letter / आवरण पत्र	Prose / memo sheets, single wide column
summary-table	Summary / सारांश	Per-title tables with totals
plain-grid	Plain grid / सादा ग्रिड	Light grid, no conversion
auto	Auto-detect / स्वतः पहचान	Uses the profiler's guess

layouts.json

Exactly a serialised WorkbookConfig — edit it, commit it, and share a printing recipe:

{
  "version": 1,
  "source": "FINAL Budget Anuman 31 October 2026.xlsx",
  "defaults": { "layout_id": "auto", "text": { "kruti": "auto" } },
  "sheets": {
    "Details":  { "layout_id": "budget-details",
                  "columns": { "mode": "explicit", "exclude": ["M", "N", "O", "P", "Q"] } },
    "Letters1": { "layout_id": "cover-letter" },
    "Saransh":  { "layout_id": "summary-table" },
    "FORMAT":   { "layout_id": "plain-grid", "text": { "kruti": "off" } }
  }
}

Key SheetConfig knobs: columns.mode (visible | native | explicit), columns.exclude / include / order / widths, rows.start / end, page_breaks.strategy (signature | title | n_rows | continuous), text.kruti (auto | on | off), text.force_english, page.orientation / margin_mm / repeat_header.

There are no environment variables — configuration lives in CLI flags and layouts.json.

REST API (app.py)

Method + path	Purpose
POST /api/workbook	Upload a workbook → workbook_id + per-sheet profiles
GET /api/workbook/{id}	Re-read profiles
PUT /api/workbook/{id}/sheets/{name}	Save one sheet's config → resolved plan + page estimate
POST /api/workbook/{id}/preview/{name}	Page-1 PNG (base64) for the drawer
POST /api/workbook/{id}/render	Render selected sheets
GET /download/{id}/{file}.pdf	Download a rendered PDF
GET /api/workbook/{id}/layouts.json	Saved WorkbookConfig
GET /api/layouts	Preset catalogue (dropdown data)

---

🛠 Tech Stack

Layer	Choice
Language	Python 3.9+
PDF engine	ReportLab (LongTable, BaseDocTemplate, custom paged canvas)
Excel I/O	openpyxl (.xlsx) · xlrd 1.2.0 (legacy .xls)
Web / API	FastAPI + Uvicorn
Thumbnails	PyMuPDF (fitz) → PNG
Front end	Vanilla JS + CSS — no build step, no npm
Fonts	Noto Sans Devanagari (Unicode) + built-in KrutiDev→Unicode transliteration

Architecture

reader → profiler → layouts (registry + resolver) → renderer
   │         │              │                          │
   └──────── schema.py (SheetConfig / WorkbookConfig / layouts.json) ───────┘
                              │
                      apply.py  ← single source of truth for
             register_fonts · make_styles · krutidev_to_unicode ·
             col_widths · build_table · segment · _PagedCanvas

apply.py is the proven legacy CLI and is imported, never modified, by the new engine/ layer — one engine, two entry points (web + CLI). See MASTER-PLAN.md for the full design.

---

🗺 Roadmap

• Dedicated prose and totals render themes (today they share the ruled path).
• Hidden-column detection for legacy .xls (xlrd colinfo_map) — currently a stub.
• Drag-to-reorder column persistence in the config drawer.
• Fold apply.py in as engine/renderer.py::cli() and retire the duplicate CLI.
• Job queue + auth for multi-user deployments.
• Word (.docx) export alongside PDF.
• Optional OCR pass for scanned annexures.

---

🤝 How to Contribute

Contributions are warmly welcome — bug reports, layout presets, docs, and translations alike.

1. Fork → create a branch: git checkout -b feat/my-idea
2. Keep changes focused; add a note to the README/roadmap if user-visible.
3. Open a Pull Request describing what and why.

Please read CONTRIBUTING.md and our Code of Conduct first. Found a security issue? See SECURITY.md.

---

📜 License

Released under the MIT License — see LICENSE.

---

❤️ Shout-outs

• PWD Tools Suite — initiative of Mrs. Premlata Jain, AAO, PWD Udaipur; the warm cream UI palette is shared with our printed budgets.
• Walkman Chanakya 905 — the public MIT KrutiDev→Unicode conversion table this project's transliteration is ported from.
• ReportLab, openpyxl, FastAPI, and Noto Sans Devanagari — the shoulders this stands on.

---

⭐ If this saves you a day of copy-paste misery — star the repo & share it!

Every star helps another district office find it. 🚀