paths.py

"""
paths.py — persistent storage layout on the Hugging Face storage bucket (/data).

If /data exists and is writable (the Space has a storage bucket attached),
everything that should survive restarts/sleep lives there:

    /data/cache_us     market-data parquet cache (yfinance downloads)
    /data/output       signal CSVs, holdings.txt
    /data/reports      auto-generated research reports (markdown)
    /data/traces       agent traces (JSON) — shareable on the Hub (📡 badge)
    /data/hf_cache     GGUF model files (downloaded once, kept forever)
    /data/pylibs       llama-cpp-python installed once at runtime, persisted

Without a bucket the app still works — it just falls back to the container
filesystem (wiped on restart).
"""
from __future__ import annotations

import os
import sys


def _writable(p: str) -> bool:
    try:
        return os.path.isdir(p) and os.access(p, os.W_OK)
    except OSError:
        return False


DATA_ROOT = "/data" if _writable("/data") else "."
PERSISTENT = DATA_ROOT == "/data"

CACHE_DIR = os.path.join(DATA_ROOT, "cache_us")
OUTPUT_DIR = os.path.join(DATA_ROOT, "output") if PERSISTENT else "./_app_output"
REPORTS_DIR = os.path.join(DATA_ROOT, "reports")
TRACES_DIR = os.path.join(DATA_ROOT, "traces")
DATASET_DIR = os.path.join(DATA_ROOT, "dataset")   # SFT training pairs (JSONL)
HF_CACHE_DIR = os.path.join(DATA_ROOT, "hf_cache")
PYLIBS_DIR = os.path.join(DATA_ROOT, "pylibs")

for _d in (CACHE_DIR, OUTPUT_DIR, REPORTS_DIR, TRACES_DIR, DATASET_DIR, HF_CACHE_DIR, PYLIBS_DIR):
    try:
        os.makedirs(_d, exist_ok=True)
    except OSError:
        pass

# GGUF downloads (hf_hub_download) go to the bucket so models persist
if PERSISTENT:
    os.environ.setdefault("HF_HOME", HF_CACHE_DIR)

# llama-cpp-python installed into the bucket must be importable.
# APPEND (not insert) so bucket-installed copies of common deps (numpy etc.)
# can never shadow the system site-packages the app was built against.
if PERSISTENT and PYLIBS_DIR not in sys.path:
    sys.path.append(PYLIBS_DIR)


def storage_status() -> str:
    if PERSISTENT:
        return ("✅ Persistent storage bucket mounted at `/data` — data cache, "
                "model files, reports, traces and the llama.cpp runtime all "
                "survive restarts.")
    return ("⚠️ No `/data` bucket detected — running on ephemeral container "
            "storage (everything re-downloads after a restart).")