feat: scaffold Week 3 assignment — validated ingestion pipeline

[lassebenni]

Summary

• Scaffolds the Week 3 assignment repo with a flat project structure that matches the assignment chapter's task numbering exactly (no task-1/task-2/ confusion)
• Auto-grader (test.sh) covers the full scoring ladder: 10 → 20 → 40 → 50 → 70 pts for the pipeline + 15 pts Azure + 15 pts AI debug
• Verified end-to-end: working solution scores 80/100, bare scaffold scores 20/100

What's included

• models.py, ingest_api.py, ingest_files.py, validate.py, database.py, pipeline.py — student starters with raise NotImplementedError TODOs
• data/weather_stations.csv — messy dataset with 6 intentional validation failures and 4 valid rows (including a duplicate that exercises the upsert path)
• output/azure_compare.md — blank template (students must write >1200 chars to earn 15 pts)
• AI_DEBUG.md — four-section template at the root (students must write >1800 chars for full marks)
• .hyf/test.sh — autograder with incremental scoring, idempotency gate, and code-pattern introspection
• .devcontainer/devcontainer.json — Python 3.11 + Azure CLI pre-installed
• AZURE_LOGIN.md — Codespaces login guide for Task 7
• .github/workflows/grade-assignment.yml — CI that runs the grader on every PR

Test plan

• Bare scaffold: 20/100 (files exist + AI_DEBUG sections), pass=false
• Working solution: 80/100 (70/70 pipeline + 10/15 AI_DEBUG), pass=true
• Idempotency: second pipeline run leaves the same row count
• Code patterns: @field_validator, @classmethod, parameterized ?, ON CONFLICT, time.sleep all detected
• Azure task: requires az login, correctly scores 0 without credentials

🤖 Generated with Claude Code

[Copilot]

Pull request overview

This PR scaffolds the Week 3 validated ingestion pipeline assignment as a flat root-level project and updates the grader/docs to match that structure.

Changes:

• Adds starter Python modules for API ingestion, CSV ingestion, validation, SQLite storage, and orchestration.
• Adds assignment templates/data for Azure comparison and AI debugging.
• Updates the autograder, README paths, devcontainer install path, and ignore rules for the root-level layout.

Reviewed changes

Copilot reviewed 9 out of 15 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
models.py	Adds the starter WeatherReading Pydantic model.
ingest_api.py	Adds API fetch/retry starter functions.
ingest_files.py	Adds CSV ingestion starter function.
validate.py	Adds batch validation starter function.
database.py	Adds SQLite helper starter functions.
pipeline.py	Adds the orchestration scaffold and expected pipeline steps.
data/weather_stations.csv	Adds the messy input dataset.
output/azure_compare.md	Adds the Azure comparison response template.
AI_DEBUG.md	Adds the AI debugging report template.
.hyf/test.sh	Updates grading paths and scoring checks for the flat structure.
.gitignore	Updates generated artifact ignore paths for the flat structure.
README.md	Updates assignment layout and local grader instructions.
requirements.txt	Adds Python dependencies.
.env.example	Adds environment variable guidance.
.devcontainer/devcontainer.json	Updates devcontainer dependency installation path.
AZURE_LOGIN.md	Provides Azure login guidance.
.github/workflows/grade-assignment.yml	Provides the grading workflow entry point.

Comments suppressed due to low confidence (2)

models.py:13

• With Pydantic v2, a @field_validator without mode="before" runs after the min_length=1 constraint. If students implement the TODO as “strip and title-case,” a whitespace-only station like " " can pass the length check first and then be stored as an empty string after stripping.
  pipeline.py:24
• This step tells students to validate “all records” together, but validate_records accepts a single source value for every error it returns. If API and CSV rows are combined before validation, the error report cannot accurately identify whether each failed row came from api or csv; the instructions should make it explicit to validate each source separately or attach source per record.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.