From 740350eacdbb12e7976d7f47e5874c58b3763d3f Mon Sep 17 00:00:00 2001
From: Marty Oehme <marty.oehme@gmail.com>
Date: Mon, 15 Jul 2024 20:06:48 +0200
Subject: [PATCH] chore(code): Refactor data processing into module

---
 pyproject.toml                                     | 2 +-
 src/{ => extract}/load_data.py                     | 2 +-
 src/{matrix.py => extract/raw_to_extracted_csv.py} | 9 ++++++---
 src/{ => extract}/yml.py                           | 0
 4 files changed, 8 insertions(+), 5 deletions(-)
 rename src/{ => extract}/load_data.py (95%)
 rename src/{matrix.py => extract/raw_to_extracted_csv.py} (81%)
 rename src/{ => extract}/yml.py (100%)

diff --git a/pyproject.toml b/pyproject.toml
index 4ad578a..60e8767 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -43,7 +43,7 @@ cmd = "nvim"
 [tool.poe.tasks.extract]
 help = "Extract the csv data from raw yaml files"
 shell = """
-python src/matrix.py > 02-data/processed/extracted.csv
+python src/extract/raw_to_extracted_csv.py > 02-data/processed/extracted.csv
 """
 [tool.poe.tasks.milestone]
 help = "Extract, render, commit and version a finished artifact"
diff --git a/src/load_data.py b/src/extract/load_data.py
similarity index 95%
rename from src/load_data.py
rename to src/extract/load_data.py
index 7cb5d8f..fe41d36 100644
--- a/src/load_data.py
+++ b/src/extract/load_data.py
@@ -5,7 +5,7 @@ from pathlib import Path
 from pandas import DataFrame, read_csv
 
 try:
-    import src.yml as yaml  # for quarto document scripts
+    import src.extract.yml as yaml  # for quarto document scripts
 except ModuleNotFoundError:
     import yml as yaml  # for directly running the package
 
diff --git a/src/matrix.py b/src/extract/raw_to_extracted_csv.py
similarity index 81%
rename from src/matrix.py
rename to src/extract/raw_to_extracted_csv.py
index d4072d5..d71f1e5 100644
--- a/src/matrix.py
+++ b/src/extract/raw_to_extracted_csv.py
@@ -8,6 +8,10 @@ try:
 except ModuleNotFoundError:
     from model import validity  # for directly running the package
 
+try:
+    import src.process.load_data as load
+except ModuleNotFoundError:
+    import load_data as load
 
 def extract(df: DataFrame, file: Path | StringIO) -> None:
     (
@@ -21,12 +25,11 @@ if __name__ == "__main__":
     import os
     import sys
 
-    import load_data
 
     if len(sys.argv) == 2:
-        df = load_data.from_yml(Path(sys.argv[1]))
+        df = load.from_yml(Path(sys.argv[1]))
     else:
-        df = load_data.from_yml()
+        df = load.from_yml()
 
     output = StringIO()
     extract(df, output)
diff --git a/src/yml.py b/src/extract/yml.py
similarity index 100%
rename from src/yml.py
rename to src/extract/yml.py