FEnigma/tests/test_ocr.py
Dominik Roth 4b427e5b0d Add a real pytest suite: 21 tests covering every OCR format + solver geometry
No test suite existed before this, which is exactly how a real
regression (the bold-span coordinate-squashing bug, and the
'Type#id:'/'<ref>: <value>' header collision, both from this session)
went unnoticed until manually re-triggered. One test per format,
cross-referenced against the full commit history so nothing already
shipped gets silently dropped by a future change:

tests/test_ocr.py: standard blocks, the calibration target line,
destruction reports (digit and letter id), train-arrival intel,
ad-hoc Enemy installations (+ their destroyed reports), Listening
Post/Coastal Battery, Marine Garrison fire-support requests, multi-
word RP names, bare-name-header targets, the bold-span coordinate-
squashing regression specifically, forward-observer report
triangulation, the '<ref>: <value>' clue grammar (+ its header-
collision regression specifically), 16-point compass tolerance, and
grid-only coordinates.

tests/test_solver.py: direct bearing+distance resolution, two-bearing
triangulation, genuine two-distance ambiguity, the nested-circles
compromise-point fallback, toleranced bearings never being used to
triangulate, and manual coord overrides clearing a stale note.

Runs via ============================= test session starts ==============================
platform linux -- Python 3.14.6, pytest-8.4.2, pluggy-1.6.0
rootdir: /home/dodox/Projects/FeNigma
configfile: pytest.ini
plugins: anyio-4.13.0
collected 21 items

tests/test_ocr.py ...............                                        [ 71%]
tests/test_solver.py ......                                              [100%]

============================== 21 passed in 0.72s ============================== (pythonpath configured in pytest.ini), dev-only
dependency in requirements-dev.txt so the app itself stays
dependency-light. Documented in the README.
2026-08-09 18:20:27 +02:00

217 lines
8.0 KiB
Python

"""Regression coverage for every intel-text format ocr.py understands.
Each format below was added incrementally in response to a real
screenshot/paste the user hit, and at least one of them (the '<ref>:
<value>' grammar colliding with the 'Type#id:' header shape) has already
regressed silently once because there was no test suite to catch it.
One test per format, named after what it covers, so a future change
that breaks an old format fails loudly and specifically instead of
being noticed (or not) days later.
"""
from fenigma import ocr
from fenigma.models import Coord, TargetType
def test_standard_target_and_rp_blocks():
text = """
Target#5 Spotted. 088, 12.10km from Spotter#1
.
Reference Point Alpha:
Bearing 094 from Spotter#1
Distance 13.26km from Spotter#2
.
AmmoCache#3:
Bearing 217 & Distance 10.48km from AmmoCache#2
"""
info = ocr.parse_text(text)
assert (TargetType.UNKNOWN, "5") in info.targets
raw, clues, coord, shell, requested_time = info.targets[(TargetType.UNKNOWN, "5")]
assert clues == [ocr.Clue(reference="Spotter#1", bearing_deg=88.0, distance_km=12.1)]
# "AmmoCache" is an alias for SupplyCache, both in the header and in
# a reference to an existing one.
assert (TargetType.SUPPLY_CACHE, "3") in info.targets
raw, clues, coord, shell, requested_time = info.targets[(TargetType.SUPPLY_CACHE, "3")]
assert clues == [ocr.Clue(reference="SupplyCache#2", bearing_deg=217.0, distance_km=10.48)]
assert "Alpha" in info.reference_points
def test_calibration_target_line():
text = "IRON NEST location - H3 5:5\nTARGET COORDINATES:\nTarget is at- Q4 4:2"
info = ocr.parse_text(text)
assert info.nest_coord == Coord("H", 3, 5, 5)
assert info.targets[(TargetType.UNKNOWN, "1")][2] == Coord("Q", 4, 4, 2)
def test_destroyed_reports_digit_and_letter_id():
text = "SupplyCache#2 Destroyed. Additional Requisition Granted.\nDirect Hit! HostileTank#3 Destroyed."
info = ocr.parse_text(text)
assert info.destroyed == {(TargetType.SUPPLY_CACHE, "2"), (TargetType.HOSTILE_TANK, "3")}
def test_train_arrival_intel():
text = """ARRIVAL STATION:
Valle de Mula MainStation: J6 0:4
.
Estimated arrival: T=10:16:50
.
TRACK ALIGNMENT:
Rail line runs straight. Bearing 090 from MainStation.
.
FINAL APPROACH:
Waypoint A - 6.00km from station: T=10:06:50
Waypoint B - 4.00km from station: T=10:10:10
"""
info = ocr.parse_text(text)
assert "MainStation" in info.reference_points
assert info.reference_points["MainStation"][2] == Coord("J", 6, 0, 4)
assert "Waypoint A" in info.reference_points
_, clues, coord = info.reference_points["Waypoint A"]
assert clues == [ocr.Clue(reference="MainStation", bearing_deg=90.0, distance_km=6.0)]
def test_enemy_multiword_name_becomes_its_own_target_type():
text = """Enemy Signal Station:
Distance 4.40km from Spotter#1
Distance 5.96km from Spotter#2
.
Enemy Assembly Area:
Bearing 034 from Enemy Signal Station
"""
info = ocr.parse_text(text)
assert (TargetType.ENEMY, "SignalStation") in info.targets
assert (TargetType.ENEMY, "AssemblyArea") in info.targets
_, clues, *_ = info.targets[(TargetType.ENEMY, "AssemblyArea")]
assert clues == [ocr.Clue(reference="Enemy#SignalStation", bearing_deg=34.0)]
def test_enemy_destroyed_report():
text = "Priority target Enemy Signal Station Destroyed, +25 Requisition."
info = ocr.parse_text(text)
assert (TargetType.ENEMY, "SignalStation") in info.destroyed
def test_listening_post_and_coastal_battery():
text = """Listening Post#1 at K6 7:8 audio reports on:
Coastal Battery#2:
Distance 6.28km South-East from Listening Post#1
"""
info = ocr.parse_text(text)
assert info.reference_points["ListeningPost#1"][2] == Coord("K", 6, 7, 8)
assert (TargetType.COASTAL_BATTERY, "2") in info.targets
_, clues, *_ = info.targets[(TargetType.COASTAL_BATTERY, "2")]
assert clues == [ocr.Clue(reference="ListeningPost#1", bearing_deg=135.0, distance_km=6.28)]
def test_marine_garrison_fire_support_request():
text = """Marine Garrison#1 pinned!
SMK Shells requested on J6 8:3
Requested before - T10:31:41 -
"""
info = ocr.parse_text(text)
assert (TargetType.MARINE_GARRISON, "1") in info.targets
raw, clues, coord, shell, requested_time = info.targets[(TargetType.MARINE_GARRISON, "1")]
assert coord == Coord("J", 6, 8, 3)
from fenigma.shells import Shell
assert shell is Shell.SMK
assert requested_time == "T10:31:41"
def test_multiword_rp_names_with_bold_markup():
text = """Reference Point <b>The Mole</b>:
Bearing <b>100°</b> from <b>Spotter#1</b>
Bearing <b>048°</b> from <b>Spotter#2</b>
.
Reference Point <b>Dockmaster's House</b>:
Distance <b>6.14km</b> from <b>The Mole</b>
"""
info = ocr.parse_text(text)
assert "TheMole" in info.reference_points
assert "DockmastersHouse" in info.reference_points
_, clues, _ = info.reference_points["DockmastersHouse"]
assert clues == [ocr.Clue(reference="TheMole", distance_km=6.14)]
def test_bare_name_header_becomes_a_target_not_an_rp():
text = """<b>HMS Rockingham</b>:
Distance <b>4.65km</b> from <b>Spotter#2</b>
"""
info = ocr.parse_text(text)
assert (TargetType.UNKNOWN, "HMSRockingham") in info.targets
assert not info.reference_points
def test_bold_coordinate_spans_are_not_corrupted_by_name_squashing():
"""A real regression: squashing multi-word bold spans into single
tokens (for names) once also mangled multi-word COORD spans like
'C9 7:9' into garbage ('C979'), because nothing distinguished the
two cases. Guarded here permanently."""
text = """IRON NEST - <b>A2 9:8</b>
.
OBSERVATION ASSETS:
<b>Spotter#1</b> - <b>C9 7:9</b>
<b>Spotter#2</b> - <b>E5 5:0</b>
"""
info = ocr.parse_text(text)
assert info.nest_coord == Coord("A", 2, 9, 8)
assert info.spotters == {1: Coord("C", 9, 7, 9), 2: Coord("E", 5, 5, 0)}
def test_forward_observer_reports_triangulate_without_leaving_ephemeral_entities():
text = """FO#5 Audio report on HMS Rockingham: 2.24km From I8 6:9 . . .
- - -
FO#4 Eyes on HMS Rockingham: 087° From G7 6:7 . . .
- - -
FO Eyes on HMS Rockingham: 099° From C9 1:2 . . .
"""
info = ocr.parse_text(text)
assert (TargetType.UNKNOWN, "HMSRockingham") in info.targets
_, clues, coord, *_ = info.targets[(TargetType.UNKNOWN, "HMSRockingham")]
# solved immediately via a scratch board, no Clues referencing an FO
# persist, and no FO ever leaks in as a real reference point.
assert clues == []
assert coord is not None
assert not info.reference_points
def test_ref_colon_value_clue_grammar():
"""'Spotter#2: 4.04km' has the exact same 'Word#digits:' shape as a
real block header ('AmmoCache#3:'), a real regression: it hijacked
the block before any clues could attach to the entity above it."""
text = """Enemy Assembly Area:
Spotter#2: 4.04km
Spotter#3: 298°
"""
info = ocr.parse_text(text)
assert (TargetType.ENEMY, "AssemblyArea") in info.targets
_, clues, *_ = info.targets[(TargetType.ENEMY, "AssemblyArea")]
assert clues == [
ocr.Clue(reference="Spotter#2", distance_km=4.04),
ocr.Clue(reference="Spotter#3", bearing_deg=298.0),
]
def test_compass_word_bearings_carry_a_tolerance():
text = """Enemy Field Command:
Spotter#1: West
Spotter#2: North-West
Spotter#3: North Northwest
"""
info = ocr.parse_text(text)
_, clues, *_ = info.targets[(TargetType.ENEMY, "FieldCommand")]
by_ref = {c.reference: c for c in clues}
assert by_ref["Spotter#1"].bearing_deg == 270.0
assert by_ref["Spotter#2"].bearing_deg == 315.0
assert by_ref["Spotter#3"].bearing_deg == 337.5
for clue in clues:
assert clue.bearing_tolerance_deg == 11.25
def test_grid_only_coord_no_sub_position():
text = "Enemy Signal Station:\n Reported active in grid D10"
info = ocr.parse_text(text)
_, clues, coord, *_ = info.targets[(TargetType.ENEMY, "SignalStation")]
assert coord == Coord("D", 10, 5, 5)