78 lines
2.9 KiB
Python
78 lines
2.9 KiB
Python
#!/usr/bin/env python3
|
|
import argparse
|
|
import json
|
|
from pathlib import Path
|
|
|
|
ROOT = Path(__file__).resolve().parents[2]
|
|
DEFAULT_MAP = ROOT / "routing/persona-contribution-map.json"
|
|
|
|
|
|
def recall(query, map_path=DEFAULT_MAP):
|
|
data = json.loads(Path(map_path).read_text(encoding="utf-8"))
|
|
needle = query.strip().casefold()
|
|
if not needle:
|
|
return {"status": "NO_TRUSTED_PATH", "query": query, "matches": []}
|
|
terms = [part for part in needle.replace("/", " ").split() if part]
|
|
matches = []
|
|
for item in data["contributions"]:
|
|
identifiers = [
|
|
value
|
|
for value in [item.get("id"), item.get("arrival_id"), *item.get("project_ids", [])]
|
|
if isinstance(value, str) and value
|
|
]
|
|
keywords = [value for value in item.get("keywords", []) if isinstance(value, str) and value]
|
|
searchable = [
|
|
*identifiers,
|
|
item.get("arrival_name"),
|
|
item.get("title"),
|
|
*keywords,
|
|
]
|
|
haystack = " ".join(value for value in searchable if isinstance(value, str) and value).casefold()
|
|
score = 0
|
|
reasons = []
|
|
for identifier in identifiers:
|
|
if identifier.casefold() in needle:
|
|
score += 100
|
|
reasons.append(identifier)
|
|
for keyword in keywords:
|
|
if keyword.casefold() in needle or keyword.casefold() == needle:
|
|
score += 20
|
|
reasons.append(keyword)
|
|
for term in terms:
|
|
if len(term) >= 2 and term in haystack:
|
|
score += 5
|
|
if needle in haystack:
|
|
score += 10
|
|
if score:
|
|
matches.append({
|
|
"score": score,
|
|
"contribution_id": item["id"],
|
|
"arrival": {"id": item["arrival_id"], "name": item["arrival_name"]},
|
|
"title": item["title"],
|
|
"matched_by": sorted(set(reasons)),
|
|
"paths": item["canonical_paths"],
|
|
"grants_execution_authority": False
|
|
})
|
|
matches.sort(key=lambda row: (-row["score"], row["contribution_id"]))
|
|
return {"status": "FOUND_CONFIDENT_PATH" if matches else "NO_TRUSTED_PATH", "query": query, "matches": matches}
|
|
|
|
|
|
def main():
|
|
parser = argparse.ArgumentParser(description="GLS-0231 read-only contribution route recall")
|
|
parser.add_argument("query")
|
|
parser.add_argument("--map", default=str(DEFAULT_MAP))
|
|
parser.add_argument("--json", action="store_true")
|
|
args = parser.parse_args()
|
|
result = recall(args.query, args.map)
|
|
if args.json:
|
|
print(json.dumps(result, ensure_ascii=False, indent=2))
|
|
return
|
|
print(result["status"])
|
|
for match in result["matches"]:
|
|
print(f'{match["contribution_id"]} · {match["arrival"]["name"]} · {match["title"]}')
|
|
for path in match["paths"]:
|
|
print(f" - {path}")
|
|
|
|
|
|
if __name__ == "__main__":
|
|
main()
|