CNEquity/codecov.yml at main · 2017-Java/CNEquity · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
# Vendored upstream (tdxpy, MIT). Measured for importability by test_tdx_decoupling,
# not line-exercised — same rationale, and the same file-level boundary, as ruff's
# extend-exclude and coverage's omit in pyproject.toml. `_wire/__init__.py` and the
# two transaction parsers are ours, not tdxpy's, so they are not listed here.
ignore:
  - "src/cnequity/adapters/tdx_protocol/_wire/constants.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/exceptions.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/heartbeat.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/helper.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/logger.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/socket_client.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/parser/base.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/parser/raw_parser.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/parser/setup_commands.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/parser/std/get_index_bars.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/parser/std/get_security_bars.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/parser/std/get_security_count.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/parser/std/get_security_list.py"
  - "src/cnequity/adapters/tdx_protocol/_wire/parser/std/get_xdxr_info.py"

coverage:
  status:
    project:
      default:
        # The gate that matters: total coverage must not rot. Branch-mode
        # (matches pytest-cov branch=true in CI). 1% absorbs the ordinary
        # wobble from a PR that moves code between files.
        target: 80%
        threshold: 1%
        # A missing upload is an infrastructure problem, not a coverage failure.
        if_not_found: success
    patch:
      default:
        # Advisory, not blocking. Patch coverage is structurally low whenever a
        # PR adds a source adapter: those are largely
        # `except ...: logger.warning(); return []` branches around live HTTP,
        # and the suite is deliberately offline, so reaching them means mocking
        # every failure mode one at a time. Gating on it turned a 0.9% miss
        # (74.08% against 75%) into a red check and pushed toward tests that
        # chase the number rather than the risk. The percentage is still
        # reported on the PR — it is just no longer a merge blocker.
        target: 75%
        threshold: 5%
        informational: true
        if_not_found: success

# One comment, updated in place, and only when coverage actually moved.
comment:
  layout: "condensed_header, condensed_files, condensed_footer"
  behavior: default
  require_changes: true