SolarCloud — AI-Powered Solar Proposal Platform

Built for Energybae, Pune · Automates the full journey from electricity bill to a professional solar proposal in under 10 seconds.

What It Does

Upload any Indian electricity bill (PDF or image). The platform uses Google Gemini Vision AI to read it, extract every key field, compute the right solar system size, and generate a downloadable package — an Excel sizing report and a branded PDF proposal — without any manual data entry.

Live Demo Flow

Upload Bill  ──►  AI Extracts Data  ──►  Review & Edit  ──►  Download ZIP
 (PDF/Image)       (Gemini Vision)       (Live ROI Chart)    (Excel + PDF)

Screenshots

Step 1 — Upload	Step 2 — Analysis	Step 3 — Download
Drag-drop bill or browse	Dark KPI card + ROI chart	Proposal ready to download

Features

AI Extraction — Gemini 1.5 Flash Vision reads MSEDCL, Adani, Tata Power, BEST, and all Indian state utility bills
Smart Fallbacks — If a field is missing, the extractor calculates it (e.g. units = current_reading − previous_reading)
Live ROI Chart — Interactive Chart.js bar chart updates in real-time as you edit any field
KPI Dashboard — Dark performance card shows units, rate, sanctioned load, monthly bill, and recommended solar size instantly
Solar System Points — Sizing list with panel count, roof area, system cost, annual savings, and payback period
Excel Report — Two-sheet workbook: Bill Data (inputs) + Solar Calculation (formula-driven, never overwritten)
PDF Proposal — 3-page branded ReportLab PDF: cover page, system sizing table, and financial ROI table
ZIP Download — Both files bundled together in a single download
Editable Review — Every extracted field is editable before generating the report
JSON Repair — Auto-fixes truncated Gemini responses so extraction never silently fails

Tech Stack

Layer	Technology
Backend	Python 3.13 · Flask 3.1 · Flask-CORS
AI / Vision	Google Gemini 1.5 Flash (`gemini-flash-latest`)
Excel	openpyxl 3.1
PDF	ReportLab (SimpleDocTemplate + Platypus)
Frontend	Vanilla JS · Chart.js 4.4.4
Fonts	Inter (Google Fonts)
Environment	python-dotenv

Project Structure

EnergyBea_assignment/
│
├── app.py                          # Flask app — API routes & request handling
├── config.py                       # Paths, API key, solar defaults (₹55K/kW, 5h sun, etc.)
├── requirements.txt                # Python dependencies
├── start.sh                        # macOS startup script (libexpat workaround)
├── .env                            # API key (not committed)
├── .env.example                    # Template for .env
│
├── services/
│   ├── extractor.py                # Gemini Vision call · JSON repair · validation
│   ├── excel_handler.py            # Excel template creation & population
│   └── pdf_generator.py           # 3-page ReportLab PDF proposal
│
├── static/
│   ├── index.html                  # Sopanel-inspired dashboard UI (3-step wizard)
│   ├── style.css                   # Dark perf-card · yellow/green theme · sidebar layout
│   └── script.js                   # Step control · extraction · live chart · ZIP download
│
├── templates/
│   └── solar_calculator_template.xlsx   # Auto-generated Excel template
│
├── uploads/                        # Temporarily stores uploaded bills (auto-created)
└── outputs/                        # Generated Excel reports (auto-created)

Setup

Prerequisites

Python 3.10+
A Google Gemini API key — free at aistudio.google.com

1. Clone & enter the project

git clone <repo-url>
cd EnergyBea_assignment

2. Create a virtual environment

python3 -m venv venv
source venv/bin/activate        # Windows: venv\Scripts\activate

3. Install dependencies

pip install -r requirements.txt

4. Add your Gemini API key

cp .env.example .env
# Open .env and set:
# GEMINI_API_KEY=your_key_here

5. Start the server

python app.py

Open http://localhost:5001 in your browser.

macOS 26 / macOS Sequoia note — If you hit a pyexpat ImportError, use the provided startup script instead:
chmod +x start.sh && ./start.sh
This sets DYLD_LIBRARY_PATH to use Homebrew's libexpat (brew install expat required).

API Reference

`GET /api/health`

Returns server status and configuration check.

{
  "status": "ok",
  "gemini_configured": true,
  "template_exists": true,
  "timestamp": "2026-05-05T10:30:00"
}

`POST /api/extract`

Upload an electricity bill and extract data using Gemini Vision.

Request: multipart/form-data with field file (PDF, PNG, JPG, WEBP, BMP, TIFF — max 10 MB)

Response:

{
  "success": true,
  "data": {
    "consumer_name": "MR. ADITYA DESHMUKH",
    "consumer_number": "610290087432",
    "billing_period": "MARCH 2026",
    "units_consumed": 560,
    "sanctioned_load": 7.5,
    "tariff_category": "LT-II Residential",
    "total_bill_amount": 7028.24,
    "electricity_rate": 12.55,
    "meter_number": "MSED778899",
    "supply_type": "Single Phase",
    "due_date": "20-APR-2026",
    "previous_reading": 15420,
    "current_reading": 15980,
    "additional_info": "MSEDCL Pune Urban Division"
  },
  "source_file": "bill.pdf"
}

`POST /api/generate`

Generate a ZIP file containing the Excel report and PDF proposal.

Request: application/json — same fields as extraction output (user-edited)

Response: Binary ZIP download (Energybae_Solar_Proposal.zip) containing:

solar_report_NAME_TIMESTAMP.xlsx
Energybae_Solar_Proposal.pdf

# Example curl
curl -X POST http://localhost:5001/api/generate \
  -H "Content-Type: application/json" \
  -o proposal.zip \
  -d '{
    "consumer_name": "Rahul Sharma",
    "units_consumed": 320,
    "sanctioned_load": 5.0,
    "total_bill_amount": 2800,
    "electricity_rate": 8.75,
    "billing_period": "March 2026"
  }'

Excel Output Structure

Sheet 1 — Bill Data (Input cells, yellow fill)

Cell	Field
B3	Consumer Name
B4	Consumer Number
B5	Billing Period
B6	Units Consumed (kWh)
B7	Sanctioned Load (kW)
B8	Tariff Category
B9	Total Bill Amount (₹)
B10	Electricity Rate (₹/kWh)

Sheet 2 — Solar Calculation (Formula cells, green fill — never overwritten)

Output	Logic
Recommended System Size (kW)	`Daily kWh ÷ 5 peak sun hours`
Number of Panels	`System kW × 1000 ÷ 540W per panel`
Roof Area Required (sq ft)	`Panels × 20 sq ft`
Annual Generation (kWh)	`System kW × 5h × 365 days`
Annual Savings (₹)	`Units generated × electricity rate`
Govt Subsidy (₹)	`₹30K/kW ≤ 2kW · ₹18K/kW for 3rd kW`
Net Investment (₹)	`System cost − subsidy`
Payback Period (years)	`Net investment ÷ annual savings`
25-Year Savings (₹)	`Compounded at 5% tariff escalation`
CO₂ Offset (tonnes/year)	`Generation × 0.82 kg/kWh ÷ 1000`

Sizing Formulas (JS + PDF mirror the Excel logic)

recommended_kw  = ⌈ units_consumed / 120 ⌉
panel_count     = ⌈ kw × 1000 / 400 ⌉       ← 400W panels
roof_area       = kw × 100  sq.ft
system_cost     = kw × ₹55,000
annual_savings  = monthly_bill × 12 × 0.90   ← 90% bill reduction
payback_years   = system_cost / annual_savings
savings_25yr    = Σ(annual_savings × 1.03^k, k=0..24) − system_cost

Supported Utility Bills

Utility	Region
MSEDCL	Maharashtra (Pune, Nagpur, etc.)
Adani Electricity	Mumbai suburbs
Tata Power	Mumbai, Delhi
BEST	Brihanmumbai
KSEB	Kerala
BESCOM	Karnataka
TANGEDCO	Tamil Nadu
Any Indian SEB	Generic fallback

Design Decisions

Formula-safe Excel writes — excel_handler.py inspects each target cell before writing. If the cell starts with =, it is skipped entirely. Input cells are populated; formula cells are never touched.

AI-first, fallback-second — Gemini Vision handles all parsing. When a field is missing (e.g. units_consumed), the extractor automatically computes it from current_reading − previous_reading. When electricity_rate is absent, it is derived as total_bill_amount ÷ units_consumed.

Resilient JSON parsing — Gemini occasionally returns truncated JSON. _repair_json() closes open braces and retries json.loads() before raising an error. The entire extraction is also retried once on failure.

Live UI computation — The ROI chart, KPI cards, and sizing list all recompute on every field edit in the browser using the same math as the backend — so what users see during review exactly matches the generated report.

In-memory ZIP — The ZIP is built in a BytesIO buffer and streamed directly — no temporary ZIP file is written to disk.

Environment Variables

Variable	Required	Description
`GEMINI_API_KEY`	Yes	Google Gemini API key from aistudio.google.com

Solar Defaults (`config.py`)

Parameter	Value
Peak Sun Hours	5.0 hrs/day (India average)
Panel Wattage	540W
Cost per kW	₹55,000
System Life	25 years
Annual Tariff Increase	5%
Annual Degradation	0.5%
CO₂ Factor	0.82 kg/kWh (India grid)

License

Internal project — Energybae, Pune. Not for public distribution.

Navigation Menu

Repository files navigation

SolarCloud — AI-Powered Solar Proposal Platform

What It Does

Live Demo Flow

Screenshots

Features

Tech Stack

Project Structure

Setup

Prerequisites

1. Clone & enter the project

2. Create a virtual environment

3. Install dependencies

4. Add your Gemini API key

5. Start the server

API Reference

`GET /api/health`

`POST /api/extract`

`POST /api/generate`

Excel Output Structure

Sheet 1 — Bill Data (Input cells, yellow fill)

Sheet 2 — Solar Calculation (Formula cells, green fill — never overwritten)

Sizing Formulas (JS + PDF mirror the Excel logic)

Supported Utility Bills

Design Decisions

Environment Variables

Solar Defaults (`config.py`)

License

About

Releases

Packages

Contributors

Languages

Name		Name	Last commit message	Last commit date
Latest commit History 28 Commits
services		services
static		static
.gitignore		.gitignore
README.md		README.md
app.py		app.py
config.py		config.py
create_test_pdf.py		create_test_pdf.py
requirements.txt		requirements.txt
sample_msedcl_bill.pdf		sample_msedcl_bill.pdf
start.sh		start.sh
test_sample.py		test_sample.py

Folders and files

Latest commit

History

Repository files navigation

SolarCloud — AI-Powered Solar Proposal Platform

What It Does

Live Demo Flow

Screenshots

Features

Tech Stack

Project Structure

Setup

Prerequisites

1. Clone & enter the project

2. Create a virtual environment

3. Install dependencies

4. Add your Gemini API key

5. Start the server

API Reference

GET /api/health

POST /api/extract

POST /api/generate

Excel Output Structure

Sheet 1 — Bill Data (Input cells, yellow fill)

Sheet 2 — Solar Calculation (Formula cells, green fill — never overwritten)

Sizing Formulas (JS + PDF mirror the Excel logic)

Supported Utility Bills

Design Decisions

Environment Variables

Solar Defaults (config.py)

License

About

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages

`GET /api/health`

`POST /api/extract`

`POST /api/generate`

Solar Defaults (`config.py`)