# -*- coding: utf-8 -*-
"""P3-S25.1 EXTERNAL TICK INGESTION & ORACLE COMPLETION -- CONFIG.

Immutable run configuration for the external Tickstory tick CSV ingestion.

Timestamp basis (established, not invented):
  - Tickstory -> MT5 tick export. Default Tickstory/MT5 export timezone is
    UTC (GMT+00). The DATE+TIME fields are interpreted as UTC wall-clock
    seconds; offset = 0. This is the same UTC-anchored hypothesis carried
    from P3-S25 and is recorded here as the immutable run configuration
    (timestamp_basis = "utc", tz_offset_seconds = 0).

Namespace is research-only. It never writes into frozen research pipelines
and never modifies external source files.
"""
import os
import hashlib
import datetime as dt

HERE = os.path.dirname(os.path.abspath(__file__))
OUT = os.path.join(HERE, "output")
SCRATCH = os.path.join(HERE, "scratch")
for d in (OUT, SCRATCH):
    os.makedirs(d, exist_ok=True)

# ---- Data sources ----------------------------------------------------------
FILES_DIR = r"D:\TradingTerminal\HFM Metatrader 5\MQL5\Files"
TICKS_CSV = os.path.join(FILES_DIR, "XAUUSD_mt5_ticks.csv")   # PRIMARY
DUKA_DIR = os.path.join(FILES_DIR, "XAUUSD")                  # FALLBACK/secondary

# ---- Timeframes ------------------------------------------------------------
M15 = 900
M30 = 1800
H4 = 14400

# ---- Timestamp semantics (immutable) --------------------------------------
TIMESTAMP_BASIS = "utc"
TZ_OFFSET_SECONDS = 0
PARSER_VERSION = "p3_s251_parser_v1"
SCHEMA_VERSION = "p3_s251_schema_v1"
ALGORITHM_VERSION = "p3_s251_algorithm_v1"

# ---- CSV column schema (no header) ----------------------------------------
# columns: date YYYYMMDD, time HH:MM:SS, bid, ask, last, volume
DATE_LEN = 8
TIME_LEN = 8
NCOL = 6


def canonical_line(epo_us, bid, ask, last, vol):
    """Deterministic CSV row of one parsed record (for chunk_sha256_parsed).

    Prices serialised with Python repr() which is deterministic for a given
    binary float value. Timestamp is integer UTC microseconds. Volume int.
    """
    return "%d,%s,%s,%s,%d\n" % (int(epo_us), repr(float(bid)),
                                 repr(float(ask)), repr(float(last)),
                                 int(vol))


def default_chunk_rows():
    """Benchmark-selected rows per chunk (1-5M allowed; 4M chosen)."""
    return 4_000_000


def main_window():
    """Primary research / population-replay window (matches P3-S25)."""
    return "20170101", "20260101"


def source_stat():
    m = os.path.getmtime(TICKS_CSV)
    return {
        "path": TICKS_CSV,
        "size_bytes": os.path.getsize(TICKS_CSV),
        "mtime_unix": float(m),
        "mtime_iso": dt.datetime.fromtimestamp(
            m, dt.timezone.utc).isoformat(),
    }