bulk-questionnaire-upload/backend

Backend

FastAPI service that validates, parses, and stores XLSForm-compatible Excel files. Deployed on Render.

Local Setup

cd backend/
python -m venv .venv
source .venv/bin/activate      # Windows: .venv\Scripts\activate
pip install -r requirements.txt
uvicorn main:app --reload --port 8000

Swagger UI: http://localhost:8000/docs

Environment Variables

Variable	Default	Description
MONGODB_URL	—	Full MongoDB connection string
DATABASE_NAME	mform_bulk_upload	Database name
FRONTEND_URL	*	Allowed CORS origin(s), comma-separated

API Reference

POST /api/validate

Validate an Excel file's structure without storing anything.

Request: multipart/form-data — field file (.xls / .xlsx)

Response:

{
  "valid": true,
  "message": "File format is valid.",
  "sheets": [{ "name": "Forms", "exists": true, "missing_columns": [] }],
  "form_metadata": { "language": "en", "title": "My Form" },
  "questions_count": 5,
  "options_count": 15,
  "errors": [],
  "warnings": []
}

POST /api/forms/parse

Parse an Excel file and return the full tempData JSON without saving to the database.

Request: multipart/form-data — field file

Response: Array — [questionnaireResponse, formDefinition]

POST /api/upload

Parse and store one or more Excel files concurrently.

Request: multipart/form-data — field files (multiple)

Response: Array of stored form objects in tempData format.

GET /api/forms

List all stored forms.

{ "forms": [{ "id": "...", "title": "...", "language": "en", "version": "1.0", "created_at": "..." }], "count": 1 }

GET /api/forms/{form_id}

Retrieve a single form in full tempData format.

PUT /api/forms/{form_id}/update

Replace a form's data with a new Excel file. multipart/form-data — field file.

DELETE /api/forms/{form_id}

Delete one form and all its questions and options.

DELETE /api/forms

Delete all forms (bulk).

File Format

Three sheets are required:

Forms

Column	Required	Notes
Language	Yes	ISO 639-1 code (en, fr, es, …)
Title	Yes	≤ 255 characters
Version	No	Arbitrary string
(others)	No	Extracted dynamically with defaults

Questions Info

Column	Required	Notes
Order	Yes	Positive integer, unique
Title	Yes	≤ 1000 characters
View Sequence	Yes	Positive integer
Input Type	Yes	1–10 (see below)

Answer Options

Column	Required	Notes
Order	Yes	References a question Order
Id	Yes	Positive integer, unique per Order
Label	Yes	≤ 500 characters

Question Types

Code	Type
1	text
2	select_one
3	select_multiple
4	integer
5	decimal
6	date
7	time
8	datetime
9	note
10	calculate

Choice questions (types 2, 3) must have matching Answer Options rows. Orphaned options (no matching question) are a validation error.

Database Schema

forms { _id, title, language, version, created_at }
questions { _id, form_id, order, title, view_sequence, input_type, created_at }
options { _id, form_id, order, option_id, label, created_at }

Indexes: forms.created_at desc, questions.form_id, options.{form_id, order}.

Testing

# Unit + integration (mocked MongoDB)
pytest tests/backend -q

# Against live deployment (saves perf_results.json)
BACKEND_URL=https://bulk-questionnaire-upload.onrender.com python scripts/measure_perf.py

Test fixtures: tests/test_xlsforms_valid/ (10 valid files) · tests/test_xlsforms_incorrect/ (error scenarios).