SniperGold_ML/tests/golden/common.py

"""Golden corpus layer definitions (spec 19, G01-G17).

Case inputs live in tests/golden/cases/*.txt (committed; the repo ignores
*.csv so fixtures use .txt). Expected outputs live in
tests/golden/expected/*.json. Expected content ids are computed with the
INDEPENDENT verifier (engine.verify.vparse) and anchored by hand-computed
primitive values (epoch ms / micro prices); the runner then requires
producer == verifier == expected.
"""

# Hand-computed anchors (independent truth for tiny inputs):
#   2023-01-02 03:04:05.100 UTC  -> 1672628645100 ms
#   2023-01-01 00:00:00.000 UTC  -> 1672531200000 ms
#   "100.000000" * 1e6 = 100_000_000 ; "100.000100" = 100_000_100
ANCHOR_TS = 1672628645100
ANCHOR_TS_2023_01_01 = 1672531200000
ANCHOR_BID = 100000000
ANCHOR_ASK = 100000100

GOLDEN_TIMEFRAMES = ["M1", "M5", "M15", "M30", "H1"]

# Default golden config (test-established tz offset = 0 UTC; has_volume False)
def golden_cfg(chunk_bytes=25_165_824, tz_offset=0):
    return {
        "source_tz_offset_minutes": tz_offset,
        "has_volume": False,
        "timeframes": list(GOLDEN_TIMEFRAMES),
        "chunk_bytes_nominal": chunk_bytes,
        "workload_bytes_nominal": 536_870_912,
    }


G01_ROWS = [
    "2023.01.02 03:04:05.100,100.000000,100.000100",
    "2023.01.02 03:04:05.200,100.000100,100.000200",
    "2023.01.02 03:04:05.300,100.000200,100.000300",
    "2023.01.02 03:04:05.400,100.000300,100.000400",
]

G03_ROWS = [
    "datetime,bid,ask",
    "2023.01.02 03:04:05.100,100.000000,100.000100",
    "2023.01.02 03:04:05.200,100.000100,100.000200",
]

G05_ROWS = list(G01_ROWS)  # 45 chars + LF = 46 bytes per line

G07_ROWS = [
    "2023.13.01 00:00:00.000,100.000000,100.000100",
    "2023.01.01 25:00:00.000,100.000000,100.000100",
    "garbage,100.000000,100.000100",
    "2023.01.01 00:00:00.000,100.000000,100.000100",
    "1999.12.31 23:59:59.999,100.000000,100.000100",
]

G08_ROWS = [
    "2023.01.01 00:00:00.000,abc,100.000100",
    "2023.01.01 00:00:01.000,-1.5,100.000100",
    "2023.01.01 00:00:02.000,1.23456789,100.000100",
    "2023.01.01 00:00:03.000,0.000000,100.000100",
    "2023.01.01 00:00:04.000,100.000000,100.000100",
]

G09_ROWS = [
    "2023.01.01 00:00:00.000,100.000100,100.000000",
    "2023.01.01 00:00:01.000,100.000000,100.000100",
]

G10_ROWS = [
    "2023.01.01 00:00:00.000,100.000000,100.000100",
    "2023.01.01 00:00:00.000,100.000000,100.000100",
    "2023.01.01 00:00:01.000,100.000000,100.000100",
]

G11_ROWS = [
    "2023.01.01 00:00:00.000,100.000000,100.000100",
    "2023.01.01 00:00:01.000,100.000100,100.000200",
    "2023.01.01 00:00:02.000,100.000200,100.000300",
]

G12_ROWS = [
    "2023.01.01 00:00:03.000,100.000300,100.000400",
    "2023.01.01 00:00:02.000,100.000100,100.000200",
    "2023.01.01 00:00:04.000,100.000000,100.000100",
]

G17_ROWS = [
    "1672628645100|100000000|100000100|1",
]


def case_bytes(case_id):
    """Return the case input bytes, exactly as committed."""
    if case_id == "G01":
        return ("\r\n".join(G01_ROWS) + "\r\n").encode("utf-8")
    if case_id == "G02":
        return ("\n".join(G01_ROWS) + "\n").encode("utf-8")
    if case_id == "G03":
        return ("\n".join(G03_ROWS) + "\n").encode("utf-8")
    if case_id == "G04":
        return ("\n".join(G01_ROWS[:2]) + "\n").encode("utf-8")
    if case_id in ("G05", "G13"):
        return ("\n".join(G05_ROWS) + "\n").encode("utf-8")
    if case_id == "G06":
        return ("\n".join(G05_ROWS) + "\n").encode("utf-8")
    if case_id == "G07":
        return ("\n".join(G07_ROWS) + "\n").encode("utf-8")
    if case_id == "G08":
        return ("\n".join(G08_ROWS) + "\n").encode("utf-8")
    if case_id == "G09":
        return ("\n".join(G09_ROWS) + "\n").encode("utf-8")
    if case_id == "G10":
        return ("\n".join(G10_ROWS) + "\n").encode("utf-8")
    if case_id == "G11":
        return ("\n".join(G11_ROWS) + "\n").encode("utf-8")
    if case_id == "G12":
        return ("\n".join(G12_ROWS) + "\n").encode("utf-8")
    if case_id == "G17":
        return ("\n".join(G17_ROWS) + "\n").encode("utf-8")
    raise KeyError(case_id)


def case_chunk_nominal(case_id):
    if case_id in ("G05", "G13"):
        return 46          # exact boundary on line end (G05) / multi-chunk (G13)
    if case_id == "G06":
        return 40          # boundary lands inside a line
    return 25_165_824