Give the coach kuhy's submission history
Some checks are pending
file length / file-length (push) Waiting to run
kuhy_credit drift / drift (push) Waiting to run
leetcode-coach pre-commit / pre-commit (push) Waiting to run
leetcode-coach tests / test (push) Waiting to run
md-naming / markdown naming convention (push) Waiting to run
palette drift / palette (push) Waiting to run
repo-contract / repo contract (push) Waiting to run
shared gate tests / gate unit tests (push) Waiting to run

history.py prints the guard's per-problem mirror (every status, code
cut at 1500 chars, newest 12); SKILL.md reads it at the start, for the
coach only. Read-only context, never a gate input.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Krzysztof Rudnicki 2026-11-22 17:36:33 +01:00
parent 41af578702
commit eddd1a38f8
4 changed files with 311 additions and 5 deletions

View File

@ -36,6 +36,8 @@ python3 ~/.claude/skills/leetcode/fetch_problem.py "<first token>"
never quote it to kuhy unasked. never quote it to kuhy unasked.
Default starter language is JavaScript (`--lang python3` etc. if kuhy asks). Default starter language is JavaScript (`--lang python3` etc. if kuhy asks).
Then `python3 ~/.claude/skills/leetcode/history.py "<first token>"`: his own attempts,
failed and accepted, with code. For you only: never paste it back; ask where it broke.
## 2. Open ## 2. Open
@ -214,8 +216,8 @@ the method appeared.
## 5d. Cold re-test ## 5d. Cold re-test
`[cold re-test]` in the arguments (`leetcode --retest <slug>`, due +3 and `[cold re-test]` in the arguments (`leetcode --retest <slug>`, due +3 and
+10 days after a solve) means kuhy solved this before; you have no history, +10 days after a solve) means kuhy solved this before; `history.py` has his old
so do not ask about the old code. In the opening, say plainly: **do NOT code, for you only: it is solved cold. In the opening, say plainly: **do NOT
submit on LeetCode** -- a resubmission mints a gate credit. Then coach as submit on LeetCode** -- a resubmission mints a gate credit. Then coach as
usual, requirement list first, and judge by running the statement's examples usual, requirement list first, and judge by running the statement's examples
locally with `node` (section 4). The normal exit and post-mortem (5, 5b) locally with `node` (section 4). The normal exit and post-mortem (5, 5b)
@ -234,12 +236,10 @@ Runs on **sonnet, opus or fable** as written. Haiku went in circles (~30
micro-questions on an Easy, 2026-10-18); `model:` frontmatter is ignored for micro-questions on an Easy, 2026-10-18); `model:` frontmatter is ignored for
skills (2026-09-18), so only if you are haiku, the **first line** of your skills (2026-09-18), so only if you are haiku, the **first line** of your
opening message is opening message is
> Type `/model sonnet` now — learning sessions are gated to sonnet, opus or > Type `/model sonnet` now — learning sessions are gated to sonnet, opus or
> fable; your next message is refused until you do (or send `keep model`). > fable; your next message is refused until you do (or send `keep model`).
then continue with the opening. `hooks/leetcode_model_gate_userprompt.sh` then continue with the opening (`hooks/leetcode_model_gate_userprompt.sh`).
enforces it from the next prompt on.
The right client is `leetcode <url>` (print mode: the Stop hook can still The right client is `leetcode <url>` (print mode: the Stop hook can still
withhold a reply; the TUI has already streamed it). `leetcode --local` runs withhold a reply; the TUI has already streamed it). `leetcode --local` runs

View File

@ -0,0 +1,111 @@
#!/usr/bin/env python3
"""Print kuhy's own submissions on one problem, failed and accepted, with code.
Usage: history.py <url-or-slug>
kuhy, 2026-11-22: a Claude coaching a problem "should also have access to my
history of approaches", including the ones that succeeded. leetcode-guard's
one-minute harvester mirrors every submission (status, time, code) into
``~/.local/share/leetcode_guard/history/<slug>.json`` (``_history_store`` in
leetcode-guard documents the shape). This reads that mirror, read-only, and
prints a bounded summary: the newest :data:`SHOWN` submissions, each code
cut at :data:`CODE_CHARS`. The mirror is display-only context and never a
gate input; nothing here writes.
Exit codes: 0 printed (also "no mirror yet" and "unreadable"), 1 bad input.
Stdlib only and free of package imports, like ``fetch_problem.py``: SKILL.md
runs it by path. Its stdout IS its interface (T201 exemption in pyproject.toml).
"""
from __future__ import annotations
from datetime import UTC, datetime
import json
from pathlib import Path
import re
import sys
from typing import Final
HISTORY_DIR: Final = Path.home() / ".local" / "share" / "leetcode_guard" / "history"
SHOWN: Final = 12
"""Newest submissions printed; older ones are counted, not shown."""
CODE_CHARS: Final = 1500
"""Code longer than this is cut: one huge paste must not flood the context."""
_ARGC: Final = 2
"""argv is the script and one problem."""
def slug_from(arg: str) -> str:
"""A full problem URL or a bare slug; anything else raises ``ValueError``."""
found = re.search(r"leetcode\.(?:com|cn)/problems/([a-z0-9-]+)", arg)
if found:
return found.group(1)
if re.fullmatch(r"[a-z0-9]+(?:-[a-z0-9]+)*", arg):
return arg
msg = f"not a LeetCode slug: {arg!r}"
raise ValueError(msg)
def _when(raw: object) -> str:
"""Unix seconds as a UTC minute, ``?`` when off."""
if isinstance(raw, int) and not isinstance(raw, bool):
return datetime.fromtimestamp(raw, tz=UTC).strftime("%Y-%m-%d %H:%M UTC")
return "?"
def _row(item: dict[str, object]) -> str:
"""One submission: a header line and its code block."""
stats = ", ".join(str(item.get(key, "?")) for key in ("lang", "runtime", "memory"))
head = f"### {item.get('status', '?')} - {_when(item.get('timestamp'))} - {stats}"
code = item.get("code")
if not isinstance(code, str):
return f"{head}\n(code not mirrored yet)"
cut = "\n[cut]" if len(code) > CODE_CHARS else ""
return f"{head}\n```\n{code[:CODE_CHARS]}{cut}\n```"
def render(arg: str) -> str:
"""The summary for one problem, as markdown."""
slug = slug_from(arg)
path = HISTORY_DIR / f"{slug}.json"
if not path.exists():
return f"No mirrored submissions for {slug} yet."
try:
data = json.loads(path.read_text(encoding="utf-8"))
except OSError, ValueError:
data = None
rows = data.get("submissions") if isinstance(data, dict) else None
if not isinstance(rows, list):
return f"The submission mirror for {slug} is unreadable."
items = [item for item in rows if isinstance(item, dict)]
items.sort(key=lambda item: _stamp(item.get("timestamp")), reverse=True)
accepted = sum(item.get("status") == "Accepted" for item in items)
title = f"{len(items)} submissions ({accepted} accepted)"
parts = [f"# kuhy's submissions on {slug}: {title}"]
parts += [_row(item) for item in items[:SHOWN]]
if len(items) > SHOWN:
parts.append(f"({len(items) - SHOWN} older not shown)")
return "\n\n".join(parts)
def _stamp(raw: object) -> int:
"""A sort key: the timestamp, 0 when off."""
return raw if isinstance(raw, int) and not isinstance(raw, bool) else 0
def main() -> int:
"""Print the summary for the problem named on argv."""
if len(sys.argv) != _ARGC:
print("usage: history.py <url-or-slug>", file=sys.stderr)
return 1
try:
print(render(sys.argv[1]))
except ValueError as exc:
print(str(exc), file=sys.stderr)
return 1
return 0
if __name__ == "__main__":
sys.exit(main())

View File

@ -0,0 +1,194 @@
"""Tests for ``history.py``: kuhy's own attempts on one problem, for the coach.
The guard's harvester mirrors every submission with its code into
``~/.local/share/leetcode_guard/history/<slug>.json`` (phase 2, 2026-11-22).
The coach reads it as context, read-only, bounded.
"""
from __future__ import annotations
import json
import runpy
import sys
from typing import TYPE_CHECKING
import pytest
from leetcode_coach import history
if TYPE_CHECKING:
from pathlib import Path
MIRROR = {
"slug": "two-sum",
"fetched_at": 1795287060,
"complete": True,
"next_offset": 0,
"submissions": [
{
"id": "3",
"status": "Accepted",
"lang": "python3",
"timestamp": 1795287000,
"runtime": "40 ms",
"memory": "17 MB",
"code": "class Solution:\n def twoSum(self): return seen",
},
{
"id": "2",
"status": "Wrong Answer",
"lang": "python3",
"timestamp": 1795286000,
"runtime": "N/A",
"memory": "N/A",
"code": "x" * 5000,
},
{
"id": "1",
"status": "Compile Error",
"lang": "python3",
"timestamp": 1795285000,
"runtime": "N/A",
"memory": "N/A",
},
],
}
@pytest.fixture(name="mirror_dir")
def _mirror_dir(tmp_path: Path, monkeypatch: pytest.MonkeyPatch) -> Path:
"""A scratch history dir; the real one is never read."""
folder = tmp_path / "history"
folder.mkdir()
monkeypatch.setattr(history, "HISTORY_DIR", folder)
return folder
def write(folder: Path, data: object, slug: str = "two-sum") -> None:
"""Put one mirror file down."""
(folder / f"{slug}.json").write_text(json.dumps(data), encoding="utf-8")
def test_every_attempt_is_listed_newest_first_with_code(mirror_dir: Path):
"""Failed and accepted alike, each with its status, and code where fetched."""
write(mirror_dir, MIRROR)
text = history.render("two-sum")
assert (
text.index("Accepted")
< text.index("Wrong Answer")
< text.index("Compile Error")
)
assert "return seen" in text
assert "(code not mirrored yet)" in text
assert "3 submissions (1 accepted)" in text
def test_long_code_is_cut(mirror_dir: Path):
"""One huge paste must not flood the coach's context."""
write(mirror_dir, MIRROR)
text = history.render("two-sum")
assert "x" * history.CODE_CHARS in text
assert "x" * (history.CODE_CHARS + 1) not in text
assert "[cut]" in text
def test_only_the_newest_submissions_are_shown(mirror_dir: Path):
"""A problem with 60 attempts shows the newest few and says how many more."""
rows = [
{"id": str(n), "status": "Wrong Answer", "timestamp": 1795280000 + n}
for n in range(60)
]
write(mirror_dir, {**MIRROR, "submissions": rows})
text = history.render("two-sum")
assert text.count("Wrong Answer") == history.SHOWN
assert f"{60 - history.SHOWN} older not shown" in text
def test_no_mirror_says_so(mirror_dir: Path):
"""No file yet (never submitted, or the harvester has not reached it)."""
assert mirror_dir.exists()
assert history.render("two-sum") == "No mirrored submissions for two-sum yet."
@pytest.mark.parametrize("data", ["{", [], {"submissions": 3}])
def test_an_unreadable_mirror_is_reported_not_trusted(mirror_dir: Path, data: object):
"""Hand-edited or half-written: say it is unreadable, never crash."""
path = mirror_dir / "two-sum.json"
path.write_text(
data if isinstance(data, str) else json.dumps(data), encoding="utf-8"
)
assert (
history.render("two-sum") == "The submission mirror for two-sum is unreadable."
)
def test_a_bad_slug_is_refused(mirror_dir: Path):
"""``../x`` is not a slug and never becomes a path."""
assert mirror_dir.exists()
with pytest.raises(ValueError, match="not a LeetCode slug"):
history.render("../escape")
def test_cli_prints_and_takes_a_url(
mirror_dir: Path,
monkeypatch: pytest.MonkeyPatch,
capsys: pytest.CaptureFixture[str],
):
"""Run by path like fetch_problem.py: a URL or a slug, exit 0."""
write(mirror_dir, MIRROR)
monkeypatch.setattr(
sys,
"argv",
["history.py", "https://leetcode.com/problems/two-sum/description/"],
)
assert history.main() == 0
assert "Accepted" in capsys.readouterr().out
def test_cli_rejects_garbage(
mirror_dir: Path,
monkeypatch: pytest.MonkeyPatch,
capsys: pytest.CaptureFixture[str],
):
"""Exit 1 with one stderr line, like fetch_problem.py's bad input."""
assert mirror_dir.exists()
monkeypatch.setattr(sys, "argv", ["history.py", "../x"])
assert history.main() == 1
assert "not a LeetCode slug" in capsys.readouterr().err
def test_runs_as_a_script(monkeypatch: pytest.MonkeyPatch, tmp_path: Path):
"""``python3 history.py <slug>`` works without the package installed."""
monkeypatch.setenv("HOME", str(tmp_path))
monkeypatch.setattr(sys, "argv", ["history.py", "two-sum"])
with pytest.raises(SystemExit) as stop:
runpy.run_path(history.__file__, run_name="__main__")
assert stop.value.code == 0
def test_cli_wants_exactly_one_problem(
monkeypatch: pytest.MonkeyPatch, capsys: pytest.CaptureFixture[str]
):
"""No argument: the usage line on stderr, exit 1."""
monkeypatch.setattr(sys, "argv", ["history.py"])
assert history.main() == 1
assert "usage" in capsys.readouterr().err
def test_a_row_without_a_timestamp_still_prints(mirror_dir: Path):
"""A hand-edited row missing its time shows ``?``, never crashes."""
write(mirror_dir, {**MIRROR, "submissions": [{"id": "9", "status": "Accepted"}]})
assert "### Accepted - ? - ?, ?, ?" in history.render("two-sum")

View File

@ -71,6 +71,7 @@ unfixable = []
"leetcode_coach/local_session.py" = ["T201"] "leetcode_coach/local_session.py" = ["T201"]
"leetcode_coach/make_cards.py" = ["T201"] "leetcode_coach/make_cards.py" = ["T201"]
"leetcode_coach/fetch_problem.py" = ["S310", "T201"] "leetcode_coach/fetch_problem.py" = ["S310", "T201"]
"leetcode_coach/history.py" = ["T201"]
"leetcode_coach/eval_local.py" = ["T201"] "leetcode_coach/eval_local.py" = ["T201"]
"leetcode_coach/stop_gate.py" = ["T201"] "leetcode_coach/stop_gate.py" = ["T201"]
"leetcode_coach/cluster_cli.py" = ["T201"] "leetcode_coach/cluster_cli.py" = ["T201"]