File size: 9,580 Bytes
d712cef | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 | #!/usr/bin/python3
"""
Weather_LLM_Strategy.py — TEXBASE Agricultural & Logistics Intelligence
==========================================================================
Reads the 16-day rain forecast CSV, computes rich statistics, then calls
the local LLM to produce a structured JSON strategy report.
Output → Weather_Forecast/weather_strategic_predictions.json
"""
import pandas as pd
import json
import requests
import os
from datetime import datetime
# ── Config ──────────────────────────────────────────────────────────────────
BASE_DIR = os.path.dirname(os.path.abspath(__file__))
CSV_PATH = os.path.join(BASE_DIR, "Weather_Forecast", 'pakistan_16day_detailed_rain.csv')
OUTPUT_JSON = os.path.join(BASE_DIR, "Weather_Forecast", 'weather_strategic_predictions.json')
LLM_URL = "https://unscotched-devon-interpapillary.ngrok-free.dev/generate"
SINDH_CITIES = {"Sanghar", "Hyderabad", "Ghotki", "Khairpur"}
PUNJAB_CITIES = {"Bahawalpur", "Multan", "Rahim Yar Khan", "Khanewal", "Bahawalnagar", "Lodhran"}
# ── LLM helper ───────────────────────────────────────────────────────────────
def call_llm(system_prompt, user_query):
try:
r = requests.post(LLM_URL,
json={"system_prompt": system_prompt, "query": user_query,
"max_new_tokens": 2000},
timeout=240)
if r.status_code == 200:
return r.json()["response"]
return f"LLM Error: {r.text}"
except Exception as e:
return f"LLM Connection Error: {e}"
# ── Data analysis helpers ────────────────────────────────────────────────────
def safe_float(val, default=0.0):
try:
return float(val) if str(val).strip() else default
except (ValueError, TypeError):
return default
def analyze_region(df, cities, region_name):
"""Compute rich statistics for a regional subset."""
rdf = df[df["City"].isin(cities)].copy()
if rdf.empty:
return {"region": region_name, "no_data": True}
rdf["Rain Amount (mm)"] = rdf["Rain Amount (mm)"].apply(safe_float)
rdf["Rain Chance (%)"] = rdf["Rain Chance (%)"].apply(safe_float)
rdf["Max Temp (°C)"] = rdf["Max Temp (°C)"].apply(safe_float)
rdf["Min Temp (°C)"] = rdf["Min Temp (°C)"].apply(safe_float)
rdf["Rain Duration (Hrs)"] = rdf["Rain Duration (Hrs)"].apply(safe_float)
# Peak events
peak_rain = rdf.loc[rdf["Rain Amount (mm)"].idxmax()] if len(rdf) > 0 else None
peak_heat = rdf.loc[rdf["Max Temp (°C)"].idxmax()] if len(rdf) > 0 else None
# Consecutive rainy days (Rain Chance > 50%) per city
max_consec = 0
for city, grp in rdf.groupby("City"):
grp = grp.sort_values("Date")
streak = consec = 0
for p in grp["Rain Chance (%)"]:
if p > 50:
consec += 1
streak = max(streak, consec)
else:
consec = 0
max_consec = max(max_consec, streak)
# Heatwave cluster: days >= 45°C anywhere in region
heatwave_days = int((rdf["Max Temp (°C)"] >= 45).sum())
# Per-city summaries
city_summaries = []
for city, grp in rdf.groupby("City"):
city_summaries.append({
"city": city,
"total_rain_mm": round(grp["Rain Amount (mm)"].sum(), 1),
"max_rain_chance_pct": int(grp["Rain Chance (%)"].max()),
"max_temp_c": round(grp["Max Temp (°C)"].max(), 1),
"avg_temp_c": round(grp["Max Temp (°C)"].mean(), 1),
"total_rain_hours": round(grp["Rain Duration (Hrs)"].sum(), 1),
})
return {
"region": region_name,
"total_rain_mm": round(rdf["Rain Amount (mm)"].sum(), 1),
"max_rain_chance_pct": int(rdf["Rain Chance (%)"].max()),
"max_temp_c": round(rdf["Max Temp (°C)"].max(), 1),
"min_temp_c": round(rdf["Min Temp (°C)"].min(), 1),
"avg_max_temp_c": round(rdf["Max Temp (°C)"].mean(), 1),
"max_consec_rainy_days": max_consec,
"heatwave_days_45plus": heatwave_days,
"peak_rain_event": {"city": str(peak_rain.get("City", "")),
"date": str(peak_rain.get("Date", "")),
"mm": safe_float(peak_rain.get("Rain Amount (mm)", 0))} if peak_rain is not None else None,
"peak_heat_event": {"city": str(peak_heat.get("City", "")),
"date": str(peak_heat.get("Date", "")),
"temp": safe_float(peak_heat.get("Max Temp (°C)", 0))} if peak_heat is not None else None,
"city_summaries": sorted(city_summaries, key=lambda x: x["total_rain_mm"], reverse=True),
}
def get_crop_season():
m = datetime.now().month
if m in [4, 5, 6]: return "Kharif Sowing (Apr-Jun) — price-sensitive"
if m in [9, 10, 11]: return "Ginning Season (Sep-Nov) — quality-critical"
if m in [12, 1, 2, 3]: return "Tight Supply Window (Dec-Mar)"
return "Pre-Sowing / Off-Season"
# ── Main ─────────────────────────────────────────────────────────────────────
def generate_weather_strategies():
if not os.path.exists(CSV_PATH):
print(f"Error: CSV not found at {CSV_PATH}")
return
print(f"Reading weather data from {CSV_PATH}...")
df = pd.read_csv(CSV_PATH)
sindh_stats = analyze_region(df, SINDH_CITIES, "Sindh")
punjab_stats = analyze_region(df, PUNJAB_CITIES, "Punjab")
# High-risk days across all cities
df["Rain Amount (mm)"] = df["Rain Amount (mm)"].apply(safe_float)
df["Rain Chance (%)"] = df["Rain Chance (%)"].apply(safe_float)
df["Max Temp (°C)"] = df["Max Temp (°C)"].apply(safe_float)
critical_days = df[(df["Rain Chance (%)"] > 60) | (df["Rain Amount (mm)"] > 15)][
["City", "Date", "Rain Amount (mm)", "Rain Chance (%)", "Max Temp (°C)"]
].to_dict(orient="records")
summary = {
"generated_at": datetime.now().isoformat(),
"forecast_period": f"{df['Date'].min()} to {df['Date'].max()}",
"crop_season": get_crop_season(),
"sindh": sindh_stats,
"punjab": punjab_stats,
"critical_days": critical_days[:20], # cap to avoid prompt overflow
"cities_covered": df["City"].unique().tolist(),
}
# ── LLM call ────────────────────────────────────────────────────────────
print("Requesting Weather Strategic Analysis from LLM...")
system_prompt = (
"You are an expert Agricultural & Logistics Strategist for TEXBASE, a Pakistan textile company. "
"Output ONLY a valid JSON object (no markdown, no preamble) with these exact keys: "
"crop_risk_assessment (string: High/Medium/Low + reasoning), "
"procurement_strategy (string: buy-now vs wait + quantities), "
"logistics_advisory (list of city-specific warnings), "
"operational_impacts (string: factory heat, warehouse moisture), "
"ginning_quality_risk (string), "
"14_day_predictions (list of date-event dicts with 'date', 'city', 'event', 'action'), "
"overall_risk_score (integer 0-10)."
)
user_query = (
f"CROP SEASON: {summary['crop_season']}\n\n"
f"SINDH STATS:\n{json.dumps(sindh_stats, indent=2)}\n\n"
f"PUNJAB STATS:\n{json.dumps(punjab_stats, indent=2)}\n\n"
f"HIGH-RISK DAYS (>60% rain or >15mm):\n{json.dumps(critical_days[:10], indent=2)}\n\n"
"Generate the JSON strategy now."
)
llm_raw = call_llm(system_prompt, user_query)
# Parse JSON — strip fences if present
try:
clean = llm_raw.strip()
if clean.startswith("```json"):
clean = clean[7:].rsplit("```", 1)[0].strip()
elif clean.startswith("```"):
clean = clean[3:].rsplit("```", 1)[0].strip()
strategy = json.loads(clean)
except Exception as e:
print(f"Warning: LLM did not return valid JSON ({e}). Storing raw.")
strategy = {"raw_llm_response": llm_raw}
# ── Save ─────────────────────────────────────────────────────────────────
output = {
"generated_at": summary["generated_at"],
"forecast_period": summary["forecast_period"],
"crop_season": summary["crop_season"],
"regional_stats": {"sindh": sindh_stats, "punjab": punjab_stats},
"critical_days": critical_days,
"weather_strategy": strategy,
}
with open(OUTPUT_JSON, "w", encoding="utf-8") as f:
json.dump(output, f, indent=4, ensure_ascii=False)
print(f"Done. Weather strategy saved → {OUTPUT_JSON}")
if __name__ == "__main__":
generate_weather_strategies()
|