Datasets:
Formats:
parquet
Languages:
English
Size:
10K - 100K
Tags:
agriculture
plant-disease
crop-disease
vision-language-model
retrieval-augmented-generation
memory
License:
File size: 1,353 Bytes
c8e03af | 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 | """Prompts, frozen from the runs that produced the earlier Ag-MEM results (prompts_frozen.json).
SAGE (KB only) and Ag-MEM use the identical system prompt, class list and output schema; only the evidence differs."""
from __future__ import annotations
import json
from pathlib import Path
_P = json.loads((Path(__file__).with_name("prompts_frozen.json")).read_text())
CAPTION_SYS = _P["caption_sys"]
DIAG_SYS = _P["diag_sys"]
def caption_text(crop: str) -> str:
return _P["caption_txt"].replace("soybean", crop.lower())
def diagnosis_user(crop: str, classes: list[str], packet: str) -> str:
return (f"Crop: {crop.lower()} (supplied).\n\n## Possible Classes\n" + "\n".join(f"- {c}" for c in sorted(classes))
+ "\n\n## Evidence packet\n" + packet + "\n\nClassify the image.")
def clean_prediction(parsed, classes: list[str]) -> dict:
p = parsed if isinstance(parsed, dict) else {}
pred = p.get("prediction")
if pred not in classes:
pred = {c.lower(): c for c in classes}.get(str(pred).lower(), "UNKNOWN")
return {"prediction": pred, "ranked": [c for c in (p.get("ranked_candidates") or []) if c in classes],
"cited": [str(e).strip("[]") for e in (p.get("used_evidence_ids") or [])], "confidence": p.get("confidence"),
"organ": p.get("organ"), "explanation": p.get("explanation")}
|