التصور المتقدم وتحليل الاتجاهات

import matplotlib.pyplot as plt
daily = pd.read_csv("datasets/kfs_farm_daily_2023_2024.csv", parse_dates=["date"]).set_index("date")
s23 = daily.loc["2023-05-01":"2023-09-30"]; s24 = daily.loc["2024-05-01":"2024-09-30"]
cmp = pd.DataFrame({"t_mean_2023": s23.air_temp_mean.values, "t_mean_2024": s24.air_temp_mean.values[:len(s23)],
                    "et0_2023": s23.et0_mm.values, "et0_2024": s24.et0_mm.values[:len(s23)]}, index=s23.index.strftime("%m-%d"))
print("mean temp May-Sep: 2023 =", round(s23.air_temp_mean.mean(), 2), "| 2024 =", round(s24.air_temp_mean.mean(), 2))
print("total ET0 May-Sep: 2023 =", round(s23.et0_mm.sum()), "mm | 2024 =", round(s24.et0_mm.sum()), "mm")
print("days with Tmax >= 38 C: 2023 =", int((s23.air_temp_max >= 38).sum()), "| 2024 =", int((s24.air_temp_max >= 38).sum()))
fig, ax = plt.subplots(figsize=(11, 4))
ax.plot(cmp.index, cmp.t_mean_2023.rolling(7).mean(), label="2023")
ax.plot(cmp.index, cmp.t_mean_2024.rolling(7).mean(), label="2024")
ax.set_xticks(cmp.index[::15]); ax.legend(); ax.set_title("KFS farm - 7-day mean air temperature, May-Sep")
fig.savefig("projects/ai/outputs/s19_season_compare.png", dpi=120, bbox_inches="tight")

season = daily.loc["2024-05-15":"2024-09-01"]                     # a typical summer maize season
tmax = season.air_temp_max.clip(upper=30); tmin = season.air_temp_min.clip(lower=10)
gdd = ((tmax + tmin) / 2 - 10).clip(lower=0)
gdd_cum = gdd.cumsum()
hourly = pd.read_csv("datasets/kfs_farm_hourly_2023_2024.csv", parse_dates=["time"]).set_index("time")
heat_hours = int((hourly.loc["2024-05-15":"2024-09-01", "air_temp_c"] >= 35).sum())
kc = pd.Series(0.3, index=season.index)                           # FAO-56 maize Kc curve (simplified)
kc.loc["2024-06-05":"2024-07-05"] = np.linspace(0.3, 1.2, len(kc.loc["2024-06-05":"2024-07-05"]))
kc.loc["2024-07-06":"2024-08-10"] = 1.2
kc.loc["2024-08-11":] = np.linspace(1.2, 0.6, len(kc.loc["2024-08-11":]))
etc = (season.et0_mm * kc).sum()
kpi = {"season GDD": round(gdd_cum.iloc[-1]), "date GDD >= 1400": str(gdd_cum[gdd_cum >= 1400].index.min().date()),
       "heat-stress hours (>=35 C)": heat_hours, "seasonal ETc (mm)": round(etc),
       "irrigation need (m3/feddan, 75% efficiency)": round(max(0, etc - season.precipitation_mm.sum()) * 4.2 / 0.75)}
print(pd.Series(kpi))

y = daily.loc["2024", "et0_mm"]
cal = pd.DataFrame({"month": y.index.month, "day": y.index.day, "et0": y.values}).pivot(index="month", columns="day", values="et0")
fig, ax = plt.subplots(figsize=(12, 4))
im = ax.imshow(cal, aspect="auto", cmap="YlOrRd")
ax.set_yticks(range(12)); ax.set_yticklabels(["Jan","Feb","Mar","Apr","May","Jun","Jul","Aug","Sep","Oct","Nov","Dec"])
ax.set_xlabel("day of month"); ax.set_title("KFS farm - daily ET0 2024 (mm/day)")
fig.colorbar(im, label="mm/day")
fig.savefig("projects/ai/outputs/s19_et0_calendar.png", dpi=120, bbox_inches="tight")
print("peak ET0 day:", y.idxmax().date(), round(y.max(), 2), "mm")

"""Session 19 - Moisture distribution map across node locations (IDW interpolation).

Edit NODES with the GPS coordinates you measured at the college farm (Google Maps -> long press).
Values come from the cleaned node log (session 14) - or type the latest readings by hand.
Output: outputs/s19_moisture_map.png  + outputs/s19_nodes.geojson (open it in Leaflet / QGIS)
"""
import json

import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

from common import OUT

# Approximate positions inside the KFS Faculty of Agriculture farm - replace with real GPS points.
NODES = {
    "greenhouse_01": (31.1052, 30.9431),
    "field_A": (31.1047, 30.9442),
    "field_B": (31.1041, 30.9436),
    "field_C": (31.1045, 30.9428),
}

clean_path = OUT / "node_log_clean.csv"
if clean_path.exists():
    log = pd.read_csv(clean_path)
    latest = log.dropna(subset=["soil_moisture_pct"]).groupby("node_id")["soil_moisture_pct"].mean()
else:
    latest = pd.Series({"greenhouse_01": 38.0, "field_A": 24.0, "field_B": 17.0})
latest = latest.reindex(NODES.keys()).fillna(latest.mean()).round(1)
print("Mean soil moisture per node (%):\n", latest)

lat = np.array([NODES[n][0] for n in latest.index]); lon = np.array([NODES[n][1] for n in latest.index])
val = latest.values

# Inverse Distance Weighting on a grid
gx, gy = np.meshgrid(np.linspace(lon.min() - 4e-4, lon.max() + 4e-4, 160),
                     np.linspace(lat.min() - 4e-4, lat.max() + 4e-4, 160))
d = np.sqrt((gx[..., None] - lon) ** 2 + (gy[..., None] - lat) ** 2) + 1e-9
w = 1 / d ** 2
grid = (w * val).sum(-1) / w.sum(-1)

fig, ax = plt.subplots(figsize=(7, 6))
im = ax.contourf(gx, gy, grid, levels=20, cmap="RdYlBu")
ax.scatter(lon, lat, c="k", s=40)
for n, x, y_, v in zip(latest.index, lon, lat, val):
    ax.annotate(f"{n}\n{v}%", (x, y_), textcoords="offset points", xytext=(6, 6), fontsize=8)
fig.colorbar(im, label="Soil moisture %")
ax.set_title("KFS farm - soil moisture distribution (IDW)")
ax.set_xlabel("Longitude"); ax.set_ylabel("Latitude")
fig.tight_layout(); fig.savefig(OUT / "s19_moisture_map.png", dpi=130)

geo = {"type": "FeatureCollection", "features": [
    {"type": "Feature", "properties": {"node": n, "soil_moisture_pct": float(v)},
     "geometry": {"type": "Point", "coordinates": [float(NODES[n][1]), float(NODES[n][0])]}}
    for n, v in latest.items()]}
(OUT / "s19_nodes.geojson").write_text(json.dumps(geo, indent=2))
print("Saved map + GeoJSON to", OUT)

{% extends "base.html" %}
{% block title %}خريطة المزرعة — KFS Smart Farm{% endblock %}
{% block head %}
<link rel="stylesheet" href="https://unpkg.com/leaflet@1.9.4/dist/leaflet.css">
{% endblock %}
{% block content %}
<section class="card">
    <div class="card-head">
        <h2>توزيع رطوبة التربة على عقد المزرعة</h2>
        <span class="legend"><i style="background:#c0392b"></i> جاف <i style="background:#f5a524"></i> متوسط <i style="background:#13787b"></i> رطب</span>
    </div>
    <div id="map" style="height:70vh;border-radius:12px"></div>
</section>
{% endblock %}
{% block scripts %}
<script src="https://unpkg.com/leaflet@1.9.4/dist/leaflet.js"></script>
<script src="https://unpkg.com/leaflet.heat@0.2.0/dist/leaflet-heat.js"></script>
<script>
const map = L.map('map').setView([{{ lat }}, {{ lon }}], 17);
L.tileLayer('https://server.arcgisonline.com/ArcGIS/rest/services/World_Imagery/MapServer/tile/{z}/{y}/{x}',
    {maxZoom: 20, attribution: 'Imagery © Esri'}).addTo(map);
const color = v => v == null ? '#888' : v < 30 ? '#c0392b' : v < 55 ? '#f5a524' : '#13787b';
fetch('/api/v1/latest').then(r => r.json()).then(({nodes}) => {
    const heat = [];
    nodes.filter(n => n.lat && n.lon).forEach(n => {
        L.circleMarker([n.lat, n.lon], {radius: 14, color: '#fff', weight: 2, fillColor: color(n.soil_moisture_pct), fillOpacity: .9})
            .addTo(map)
            .bindPopup(`<b>${n.name || n.node_id}</b><br>رطوبة التربة: ${n.soil_moisture_pct ?? '—'}%<br>الحرارة: ${n.air_temp_c ?? '—'} °C<br>${n.online ? '🟢 متصل' : '🔴 غير متصل'}`);
        if (n.soil_moisture_pct != null) heat.push([n.lat, n.lon, 1 - n.soil_moisture_pct / 100]);
    });
    if (heat.length) L.heatLayer(heat, {radius: 45, blur: 30, gradient: {0.2: '#13787b', 0.5: '#f5a524', 0.9: '#c0392b'}}).addTo(map);
});
</script>
{% endblock %}

صفحة /map في المنصة: صور أقمار صناعية (Esri) + دائرة لكل عقدة بلون رطوبتها + طبقة حرارية (leaflet.heat) تتحدث من /api/v1/latest. الإحداثيات من جدول nodes — حدّثها بـ GPS الحقيقي.

from matplotlib.backends.backend_pdf import PdfPages
with PdfPages("projects/ai/outputs/kfs_farm_report.pdf") as pdf:
    for title, series in [("Daily mean air temperature 2024", daily.loc["2024", "air_temp_mean"]),
                          ("Daily ET0 2024 (mm)", daily.loc["2024", "et0_mm"]),
                          ("Cumulative GDD - maize season 2024", gdd_cum)]:
        fig, ax = plt.subplots(figsize=(10, 4))
        series.plot(ax=ax); ax.set_title(title); ax.grid(alpha=.3)
        pdf.savefig(fig); plt.close(fig)
print("report saved: kfs_farm_report.pdf (3 pages)")