MF-Structure Query Generator

Classification: Team coursework for CS562. Repository history credits both E-P-C and Pangqiang-Gary; it is not presented as a solo project.

This project parses an MF-structure query specification and generates an executable Python query processor for a PostgreSQL sales relation. It makes the steps between a declarative specification, grouping-variable scans, aggregate state, the G predicate, and final output explicit.

My contribution

My public commits include the MF-structure parser and query-generator implementation, reproducible cases, an additional n=2/HAVING case, and integration work. The main implementation is visible in this commit; the full contributor history preserves the team attribution.

Pipeline

phi specification -> parser -> typed MF representation -> Python code generator -> PostgreSQL query processor

The input format contains:

• S — selected output attributes;
• n — number of grouping variables;
• V — grouping attributes;
• F — aggregate functions;
• sigma — per-grouping-variable predicates;
• G — post-aggregation condition.

Reproducible cases

The repository contains six specifications (phi_input_case1.txt through phi_input_case6.txt) and their recorded outputs under Output/. The extended implementation and additional generated processors are being reviewed in PR #2.

Quick start

python -m venv .venv
# Windows: .venv\Scripts\activate
# macOS/Linux: source .venv/bin/activate
pip install -r requirements.txt

Create local database configuration from the safe template:

# Windows
copy .env.example .env

# macOS/Linux
cp .env.example .env

Generate a processor from one case:

python codegen.py phi_input_case1.txt generated_case1.py
python generated_case1.py

The configured PostgreSQL database must contain the sales relation expected by the selected specification. load_sales_10000_table (NEW).sql provides the project dataset setup script.

Repository map

Path	Purpose
phi_parser.py	Parses and validates the MF specification
codegen.py	Emits an executable Python query processor
qpe.py	Example generated processor
phi_input_case*.txt	Six reproducible specifications
Output/case*.txt	Recorded outputs for comparison
COLAB_ONE_CELL_SETUP.txt	Self-contained Colab setup helper

Validation

CI checks every committed Python source file and generates a compilable processor for each of the six specifications without requiring database credentials.

python -m py_compile codegen.py phi_parser.py qpe.py
python codegen.py phi_input_case1.txt generated_case1.py
python -m py_compile generated_case1.py

Security and scope

Database credentials belong only in the untracked .env file. .env.example contains placeholders and is safe to commit. Generated predicates are evaluated in a restricted namespace, but specifications should still be treated as trusted project inputs rather than arbitrary user input.