from __future__ import annotations import json import os import re import subprocess from pathlib import Path from typing import Any from fastapi import FastAPI, HTTPException, Query from fastapi.responses import FileResponse from fastapi.staticfiles import StaticFiles BASE_DIR = Path(__file__).resolve().parent STATIC_DIR = BASE_DIR / "static" PGHOST = os.getenv("PGHOST", "192.168.3.3") PGPORT = os.getenv("PGPORT", "5432") PGUSER = os.getenv("PGUSER", "his_user") PGDATABASE = os.getenv("PGDATABASE", "pacs_db") PACS_TABLE = os.getenv("PACS_TABLE", "pacs_dicom_files") PACS_SUMMARY_TABLE = os.getenv("PACS_SUMMARY_TABLE", "pacs_dicom_study_summaries") UPP_ASSET_TABLE = os.getenv("UPP_ASSET_TABLE", "upp_exam_assets") UPP_STL_TABLE = os.getenv("UPP_STL_TABLE", "upp_stl_files") IDENT_RE = re.compile(r"^[A-Za-z_][A-Za-z0-9_]*$") def ident(name: str) -> str: if not IDENT_RE.fullmatch(name): raise RuntimeError(f"invalid SQL identifier: {name}") return name PACS_TABLE_SQL = ident(PACS_TABLE) PACS_SUMMARY_TABLE_SQL = ident(PACS_SUMMARY_TABLE) UPP_ASSET_TABLE_SQL = ident(UPP_ASSET_TABLE) UPP_STL_TABLE_SQL = ident(UPP_STL_TABLE) app = FastAPI(title="PACS UPP Database Visualizer") app.mount("/static", StaticFiles(directory=STATIC_DIR), name="static") def pg_env() -> dict[str, str]: env = os.environ.copy() env.update({"PGHOST": PGHOST, "PGPORT": PGPORT, "PGUSER": PGUSER, "PGDATABASE": PGDATABASE}) return env def run_psql(sql: str, timeout: int = 20) -> subprocess.CompletedProcess[str]: return subprocess.run( ["psql", "-X", "-q", "-t", "-A", "-v", "ON_ERROR_STOP=1", "-c", sql], text=True, capture_output=True, timeout=timeout, env=pg_env(), ) def pg_scalar(sql: str, timeout: int = 20) -> str: result = run_psql(sql, timeout=timeout) if result.returncode != 0: raise RuntimeError(result.stderr.strip() or result.stdout.strip()) return result.stdout.strip() def pg_json_rows(select_sql: str, timeout: int = 24) -> list[dict[str, Any]]: payload = pg_scalar(f"SELECT COALESCE(json_agg(row_to_json(q)), '[]'::json)::text FROM ({select_sql}) q", timeout=timeout) return json.loads(payload or "[]") def sql_literal(value: Any) -> str: return "'" + str(value).replace("'", "''") + "'" def normalize_ct(value: str) -> str: return re.sub(r"^D", "", re.sub(r"\s+", "", str(value or "").upper())) def db_available() -> tuple[bool, str]: try: pg_scalar("SELECT 1", timeout=4) return True, "connected" except Exception as exc: # noqa: BLE001 return False, str(exc) def relation_cte() -> str: return f""" WITH p AS ( SELECT p.*, regexp_replace(upper(p.ct_number), '^D', '') AS norm_ct FROM public.{PACS_TABLE_SQL} p ), ps AS ( SELECT * FROM public.{PACS_SUMMARY_TABLE_SQL} ), u_raw AS ( SELECT u.*, regexp_replace(upper(u.ct_number), '^D', '') AS norm_ct FROM public.{UPP_ASSET_TABLE_SQL} u ), u AS ( SELECT DISTINCT ON (norm_ct) * FROM u_raw ORDER BY norm_ct, stl_present DESC, list_present DESC, updated_at DESC NULLS LAST ), s_raw AS ( SELECT s.*, regexp_replace(upper(s.ct_number), '^D', '') AS norm_ct FROM public.{UPP_STL_TABLE_SQL} s ), s AS ( SELECT DISTINCT ON (norm_ct) * FROM s_raw ORDER BY norm_ct, file_count DESC, updated_at DESC NULLS LAST ), keys AS ( SELECT norm_ct FROM p UNION SELECT norm_ct FROM u ), relation AS ( SELECT k.norm_ct, p.ct_number AS pacs_ct_number, u.ct_number AS upp_ct_number, s.ct_number AS stl_ct_number, p.ct_number IS NOT NULL AS pacs_present, u.ct_number IS NOT NULL AS upp_present, COALESCE(u.stl_present, false) OR s.ct_number IS NOT NULL AS stl_present, CASE WHEN p.ct_number IS NOT NULL AND u.ct_number IS NOT NULL AND (COALESCE(u.stl_present, false) OR s.ct_number IS NOT NULL) THEN 'complete' WHEN p.ct_number IS NOT NULL AND u.ct_number IS NOT NULL THEN 'no_stl' WHEN p.ct_number IS NOT NULL THEN 'pacs_only' WHEN u.ct_number IS NOT NULL THEN 'upp_only' ELSE 'unknown' END AS relation_status, CASE WHEN p.ct_number IS NOT NULL AND u.ct_number IS NOT NULL AND p.ct_number = u.ct_number THEN 'exact' WHEN p.ct_number IS NOT NULL AND u.ct_number IS NOT NULL THEN 'normalized' ELSE '' END AS match_type, p.batch_name, p.source_patient_name, p.patient_name_dicom, COALESCE(NULLIF(p.source_patient_name, ''), NULLIF(p.patient_name_dicom, '')) AS pacs_patient_name, p.patient_id, p.patient_sex AS pacs_patient_sex, p.study_date, p.study_time, p.study_description, p.modality, p.series_count AS pacs_series_count, p.dicom_file_count, p.processed_path, p.updated_at AS pacs_updated_at, ps.annotated_series, ps.undetermined_series, ps.completed, ps.body_parts, u.patient_name AS upp_patient_name, u.patient_sex AS upp_patient_sex, u.patient_age, u.patient_id_masked, u.exam_date, u.task_created_at, u.exam_description, u.exam_device, u.algorithm_model, u.upp_status, u.list_present, u.list_record_count, u.processed_stl_dir, u.stl_file_count, u.stl_total_bytes, u.updated_at AS upp_updated_at, u.selected_list_record, u.list_records, u.stl_candidates, s.file_count AS stl_file_count_agg, s.total_bytes AS stl_total_bytes_agg, s.segment_names, s.segment_families, s.segment_categories, s.file_names, s.files AS stl_files, s.updated_at AS stl_updated_at, CASE WHEN COALESCE(NULLIF(p.source_patient_name, ''), NULLIF(p.patient_name_dicom, '')) IS NULL OR NULLIF(u.patient_name, '') IS NULL THEN '' WHEN regexp_replace(COALESCE(NULLIF(p.source_patient_name, ''), p.patient_name_dicom), '\\s+', '', 'g') = regexp_replace(u.patient_name, '\\s+', '', 'g') THEN 'same' ELSE 'different' END AS patient_name_match FROM keys k LEFT JOIN p ON p.norm_ct = k.norm_ct LEFT JOIN ps ON ps.ct_number = p.ct_number LEFT JOIN u ON u.norm_ct = k.norm_ct LEFT JOIN s ON s.norm_ct = k.norm_ct ) """ def relation_where(q: str, status: str) -> str: clauses: list[str] = [] query = q.strip() if query: like = "%" + query.replace("%", "").replace("_", "") + "%" literal = sql_literal(like) clauses.append( " OR ".join( [ f"norm_ct ILIKE {literal}", f"pacs_ct_number ILIKE {literal}", f"upp_ct_number ILIKE {literal}", f"pacs_patient_name ILIKE {literal}", f"upp_patient_name ILIKE {literal}", f"patient_id ILIKE {literal}", f"exam_description ILIKE {literal}", f"study_description ILIKE {literal}", ] ).join(("(", ")")) ) if status == "complete": clauses.append("relation_status = 'complete'") elif status == "no_stl": clauses.append("relation_status = 'no_stl'") elif status == "pacs_only": clauses.append("relation_status = 'pacs_only'") elif status == "upp_only": clauses.append("relation_status = 'upp_only'") elif status == "normalized": clauses.append("match_type = 'normalized'") elif status == "undetermined": clauses.append("COALESCE(undetermined_series, 0) > 0") elif status != "all": raise HTTPException(status_code=400, detail="invalid status filter") return "WHERE " + " AND ".join(clauses) if clauses else "" @app.get("/") def index() -> FileResponse: return FileResponse(STATIC_DIR / "index.html") @app.get("/health") def health() -> str: return "ok" @app.get("/api/status") def status() -> dict[str, Any]: ok, message = db_available() counts: dict[str, Any] = {} if ok: rows = pg_json_rows( f""" {relation_cte()} SELECT count(*)::int AS total_ct, count(*) FILTER (WHERE pacs_present)::int AS pacs_count, count(*) FILTER (WHERE upp_present)::int AS upp_count, count(*) FILTER (WHERE stl_present)::int AS stl_count, count(*) FILTER (WHERE relation_status = 'complete')::int AS complete_count, count(*) FILTER (WHERE relation_status = 'no_stl')::int AS no_stl_count, count(*) FILTER (WHERE relation_status = 'pacs_only')::int AS pacs_only_count, count(*) FILTER (WHERE relation_status = 'upp_only')::int AS upp_only_count, count(*) FILTER (WHERE match_type = 'normalized')::int AS normalized_match_count, COALESCE(sum(COALESCE(undetermined_series, 0)), 0)::int AS undetermined_series FROM relation """, timeout=24, ) counts = rows[0] if rows else {} return { "database": {"ok": ok, "message": message, "host": PGHOST, "database": PGDATABASE}, "tables": { "pacs": PACS_TABLE, "pacs_summary": PACS_SUMMARY_TABLE, "upp_assets": UPP_ASSET_TABLE, "upp_stl": UPP_STL_TABLE, }, "counts": counts, } @app.get("/api/relations") def relations(q: str = "", status: str = "all", limit: int = Query(default=300, ge=1, le=1000)) -> list[dict[str, Any]]: where = relation_where(q, status) return pg_json_rows( f""" {relation_cte()} SELECT norm_ct, pacs_ct_number, upp_ct_number, stl_ct_number, pacs_present, upp_present, stl_present, relation_status, match_type, pacs_patient_name, upp_patient_name, patient_name_match, patient_id, patient_id_masked, study_date, study_time, exam_date, task_created_at, study_description, exam_description, algorithm_model, upp_status, pacs_series_count, dicom_file_count, annotated_series, undetermined_series, completed, list_present, list_record_count, stl_file_count, stl_file_count_agg, body_parts, segment_categories, segment_families, pacs_updated_at, upp_updated_at, stl_updated_at FROM relation {where} ORDER BY pacs_present DESC, relation_status, COALESCE(study_date, '') DESC, COALESCE(study_time, '') DESC, norm_ct LIMIT {int(limit)} """, timeout=24, ) @app.get("/api/relations/{ct_number}") def relation_detail(ct_number: str) -> dict[str, Any]: norm_ct = normalize_ct(ct_number) if not norm_ct: raise HTTPException(status_code=400, detail="invalid CT number") rows = pg_json_rows( f""" {relation_cte()} SELECT * FROM relation WHERE norm_ct = {sql_literal(norm_ct)} LIMIT 1 """, timeout=24, ) if not rows: raise HTTPException(status_code=404, detail="CT number not found") return rows[0]