update minio port

backend/app/algo_host/job_workspace.py

@@ -0,0 +1,120 @@
+"""Prepare input artifacts expected by algorithm_subprocesses/5.15 main.py."""
+
+from __future__ import annotations
+
+import copy
+import json
+from dataclasses import dataclass
+from pathlib import Path
+
+import yaml
+
+from app.algo_host.bundle import load_reference_default_config, resolve_reference_bundle_dir
+from app.algo_host.transcode import ensure_batch_pipeline_input_video
+from app.consumable_catalog import build_name_mapping
+
+
+@dataclass(frozen=True)
+class BatchJobFiles:
+    config_path: Path
+    excel_path: Path
+    whitelist_path: Path
+    output_path: Path
+    work_dir: Path
+    input_video_path: Path
+
+
+def write_reference_catalog_excel(
+    path: Path,
+    *,
+    candidate_consumables: list[str],
+) -> None:
+    import pandas as pd
+
+    name_to_code = build_name_mapping(candidate_consumables)
+    rows = [
+        {
+            "序号": idx,
+            "产品编码": name_to_code.get(name, name),
+            "商品名称": name,
+        }
+        for idx, name in enumerate(candidate_consumables, start=1)
+    ]
+    path.parent.mkdir(parents=True, exist_ok=True)
+    pd.DataFrame(rows, columns=["序号", "产品编码", "商品名称"]).to_excel(path, index=False)
+
+
+def write_reference_whitelist_json(path: Path, *, candidate_consumables: list[str]) -> None:
+    path.parent.mkdir(parents=True, exist_ok=True)
+    path.write_text(
+        json.dumps({"allowed_names": candidate_consumables}, ensure_ascii=False, indent=2),
+        encoding="utf-8",
+    )
+
+
+def build_job_config(
+    *,
+    bundle_dir: Path,
+    video_path: Path,
+    output_path: Path,
+    work_dir: Path,
+    excel_path: Path,
+    whitelist_path: Path,
+) -> dict:
+    cfg = copy.deepcopy(load_reference_default_config(bundle_dir))
+    cfg["io"]["video"] = str(video_path.resolve())
+    cfg["io"]["excel"] = str(excel_path.resolve())
+    cfg["io"]["out"] = str(output_path.resolve())
+    cfg["io"]["whitelist_json"] = str(whitelist_path.resolve())
+    cfg["runtime"]["work_dir"] = str(work_dir.resolve())
+    cfg["runtime"]["keep_work_dir"] = False
+    return cfg
+
+
+def prepare_batch_job(
+    *,
+    bundle_dir: Path | None,
+    cache_dir: Path,
+    uploaded_video_path: Path,
+    candidate_consumables: list[str],
+) -> BatchJobFiles:
+    root = resolve_reference_bundle_dir(bundle_dir)
+    cache_input_dir = cache_dir / "input"
+    cache_output_dir = cache_dir / "output"
+    cache_work_dir = cache_dir / "work"
+    cache_config_dir = cache_dir / "config"
+    for d in (cache_input_dir, cache_output_dir, cache_work_dir, cache_config_dir):
+        d.mkdir(parents=True, exist_ok=True)
+
+    cache_input = cache_input_dir / "input.mp4"
+    ensure_batch_pipeline_input_video(
+        source_path=uploaded_video_path,
+        dest_path=cache_input,
+    )
+    output_path = cache_output_dir / "result.tsv"
+    excel_path = cache_config_dir / "商品信息表.xlsx"
+    whitelist_path = cache_config_dir / "whitelist.json"
+    config_path = cache_config_dir / "config.yaml"
+
+    write_reference_catalog_excel(excel_path, candidate_consumables=candidate_consumables)
+    write_reference_whitelist_json(whitelist_path, candidate_consumables=candidate_consumables)
+    config = build_job_config(
+        bundle_dir=root,
+        video_path=cache_input.resolve(),
+        output_path=output_path.resolve(),
+        work_dir=cache_work_dir.resolve(),
+        excel_path=excel_path.resolve(),
+        whitelist_path=whitelist_path.resolve(),
+    )
+    config_path.write_text(
+        yaml.safe_dump(config, allow_unicode=True, sort_keys=False),
+        encoding="utf-8",
+    )
+    return BatchJobFiles(
+        config_path=config_path,
+        excel_path=excel_path,
+        whitelist_path=whitelist_path,
+        output_path=output_path,
+        work_dir=cache_work_dir,
+        input_video_path=cache_input,
+    )