"""Build the bilingual, static Food analysis from the public project sources.
Run the existing reviewed_analysis.py first to refresh results/metrics.json.
Then run this file. It preserves the case-study header and brand assets,
computes descriptive tables, and renders the archived notebook without running it.
No data, code, or account is fetched by the published pages.
"""
import html
import json
import re
from pathlib import Path
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
HERE = Path(__file__).resolve().parent
ROOT = HERE.parent.parent
RESULTS = HERE / "results"
METRICS = json.loads((RESULTS / "metrics.json").read_text())
RAW = pd.read_csv(HERE / "nutrition.csv")
NAMES = {"Calories": "calories", "Protein": "protein", "Carbohydrate": "carb",
"Total fat": "fat", "Cholesterol": "cholesterol", "Fiber": "fiber",
"Water": "water", "Alcohol": "alcohol", "Vitamin C": "vitamin_c"}
DATA = RAW[["FDC_ID", "Item", "Category"]].copy()
for source, target in NAMES.items():
DATA[target] = pd.to_numeric(RAW[source].astype("string").str.extract(
r"^\s*([-+]?\d+(?:\.\d+)?)", expand=False), errors="coerce")
# Stop instead of publishing tables beside stale headline metrics.
assert len(DATA) == METRICS["rows"]
assert DATA.Category.nunique() == METRICS["categories"]
assert DATA[list(NAMES.values())].isna().sum().to_dict() == METRICS["missing"]
assert np.isclose(DATA[["water", "calories"]].corr().iloc[0, 1],
METRICS["waterCaloriesPearsonR"])
FRUITS = DATA[DATA.Category.eq("Fruits and Fruit Juices")].dropna(
subset=["vitamin_c"]).nlargest(5, "vitamin_c")
assert FRUITS[["Item", "vitamin_c"]].to_dict("records") == METRICS["topFruitVitaminC"]
CATEGORIES = DATA.groupby("Category", sort=True).size().rename("foods").reset_index()
CATEGORIES.to_csv(RESULTS / "category_counts.csv", index=False)
FRUITS[["FDC_ID", "Item", "vitamin_c"]].to_csv(RESULTS / "fruit_vitamin_c.csv", index=False)
GROUPS = [("All foods", DATA), ("Zero-carbohydrate foods", DATA[DATA.carb.eq(0)]),
("500 highest-protein foods", DATA.nlargest(500, "protein"))]
GROUP_RESULTS = []
for name, rows in GROUPS:
GROUP_RESULTS.append({"group": name, "rows": len(rows),
"cholesterol_rows": int(rows.cholesterol.notna().sum()),
"mean_calories_kcal": float(rows.calories.mean()),
"mean_fat_g": float(rows.fat.mean()),
"mean_cholesterol_mg": float(rows.cholesterol.mean())})
pd.DataFrame(GROUP_RESULTS).to_csv(RESULTS / "food_group_comparison.csv", index=False)
plt.rcParams.update({"font.family": "DejaVu Sans", "font.size": 11,
"axes.spines.top": False, "axes.spines.right": False,
"axes.spines.left": False, "axes.edgecolor": "#d7ddd7",
"text.color": "#102a43", "axes.labelcolor": "#60706a",
"xtick.color": "#60706a", "ytick.color": "#102a43",
"svg.hashsalt": "shams-food-study-v1"})
def chart(path, labels, values, xlabel, maximum, colors):
"""Generate exact vector charts; labels remain selectable text."""
plt.rcParams["svg.fonttype"] = "none"
fig, ax = plt.subplots(figsize=(10, 4.9), layout="constrained")
fig.patch.set_facecolor("#fffdf7")
ax.set_facecolor("#fffdf7")
y = np.arange(len(values))
ax.barh(y, values, color=colors, height=.55, zorder=3)
ax.set_yticks(y, labels)
ax.invert_yaxis()
ax.set_xlim(0, maximum)
ax.set_xlabel(xlabel, labelpad=14)
ax.tick_params(axis="y", length=0, pad=12)
ax.grid(axis="x", color="#d7ddd7", alpha=.7, zorder=0)
for pos, value in zip(y, values):
ax.text(value + maximum * .015, pos, f"{value:,.2f}" if maximum < 100 else f"{value:,.1f}",
va="center", fontsize=11, fontweight="bold")
fig.savefig(path, metadata={"Date": None})
plt.close(fig)
chart(RESULTS / "fruit-vitamin-c.svg",
["Acerola, raw", "Acerola juice, raw", "Guavas, raw",
"Jujube, fresh, dried*", "Litchis, dried"], FRUITS.vitamin_c.tolist(),
"Vitamin C (mg per 100 g food)", 2000,
["#00766c", "#5b9f93", "#5b9f93", "#5b9f93", "#5b9f93"])
chart(RESULTS / "model-errors.svg",
["Macronutrients", "Category\ninteractions", "Alcohol + fiber\ninteractions"],
[m["test_rmse_kcal"] for m in METRICS["models"]],
"Test RMSE (kcal per 100 g food) · lower is better", 19,
["#6f8c94", "#6f8c94", "#00766c"])
def esc(value):
return html.escape(str(value), quote=True)
def table(headers, rows, caption=None):
cap = f"
{esc(caption)}
" if caption else ""
head = "".join(f'
{esc(h)}
' for h in headers)
body = []
for row in rows:
cells = []
for item in row:
numeric = ' data-number=""' if re.fullmatch(r"[-\d.,% /]+", str(item)) else ""
cells.append(f"{esc(item)}")
body.append("
" + "".join(cells) + "
")
return f'
{cap}
{head}
{"".join(body)}
'
def details(label, body, code=False):
content = f'
{esc(body)}
' if code else f'
{body}
'
return f'{esc(label)}{content}'
def archive_text(source):
"""Render only the simple, observed markdown subset; never include raw HTML."""
safe = esc(source)
safe = re.sub(r"\[([^\]]+)\]\((https?://[^\s)]+)\)",
r'\1', safe)
safe = re.sub(r"\*\*([^*]+)\*\*", r"\1", safe)
safe = re.sub(r"`([^`]+)`", r"\1", safe)
safe = re.sub(r"^#{1,6}\s+(.+)$", r"
')
ARCHIVE_HTML = "".join(ARCHIVE)
SOURCE = (HERE / "reviewed_analysis.py").read_text()
VITAMIN_SOURCE = (HERE / "vitamin_c_analysis.py").read_text()
def build(lang):
ar = lang == "ar"
t = lambda en, arabic: arabic if ar else en
suffix = "-ar" if ar else ""
filename = f"nutrition-analysis{suffix}.html"
base = (ROOT / f"nutrition{suffix}.html").read_text()
# Reuse the current edition's header, language control, fonts and footer.
head = base.split("", 1)[0]
title = t("Vitamin C and calorie analysis, results and code | Shams Insights", "تحليل فيتامين C والسعرات والنتائج والكود | شمس إنسايتس")
desc = t("Read the complete Food study: data cleaning, vitamin C, water and energy, regression results, reproducible Python and the archived 2023 notebook. No sign-in.",
"اقرأ تحليل الأغذية كاملًا: تنظيف البيانات وفيتامين C والماء والسعرات ونتائج الانحدار وكود Python ودفتر 2023 الأصلي، دون تسجيل دخول.")
head = re.sub(r".*?", f"{title}", head)
head = re.sub(r'', f'', head)
head = re.sub(r'', f'', head)
head = re.sub(r'', f'', head)
head = head.replace("nutrition-ar.html", "nutrition-analysis-ar.html").replace("nutrition.html", "nutrition-analysis.html")
if "nutrition-analysis.css" not in head:
head += ''
header = re.search(r'()', base, re.S).group(1)
header = header.replace('href="nutrition.html" lang="en"', 'href="nutrition-analysis.html" lang="en"')
header = header.replace('href="nutrition-ar.html" lang="ar"', 'href="nutrition-analysis-ar.html" lang="ar"')
footer = base.split("", 1)[1]
footer = re.sub(r'