mirror of
https://github.com/zvx-echo6/meshai.git
synced 2026-08-26 17:31:34 +00:00
fix(nws): tighten hazard wording across all product types; kill dangling '— …' in L4
Special Weather Statements (and other SPS/WSW/FFW/FLW products) rendered a
verbose raw hazard sentence on L3 that ate the packet budget, collapsing L4 to
a dangling 'Moving SW 24 mph —…' with every town lost (seen live in the
Activity Log).
- Add _tighten_hazard(): compacts free-form NWS hazard text into the terse SVR
idiom for ALL branches — 'Wind gusts in excess of 45 mph' -> '45mph gusts',
'in excess of' -> '>', '45 mph' -> '45mph', 'pea size hail' -> '0.25" hail'.
Applied to the FFW/FLW and SPS/WSW/else branches (SVR already terse).
- Rework L4 assembly to be budget-aware BEFORE the final hard cap: try location
forms richest->poorest (full list -> first->mid->last -> first->last ->
first-only -> none) and only attach '— {locs}' when the whole message fits.
If no location fits, degrade to motion-only; if even that overflows, drop L4.
A dangling '— …' / trailing '—' is now structurally impossible.
- Tests: SPS worst-case (tightened + no dangling), WSW, pathological
motion-only degrade, SVR no-dangling re-verify; shared dangling-separator
assertion.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
Matt Johnson
parent
0460462485
commit
5381975e83
2 changed files with 249 additions and 28 deletions
|
|
@ -374,6 +374,45 @@ def _collapse_certainty(text: str) -> str:
|
|||
return (text or "").strip()
|
||||
|
||||
|
||||
def _tighten_hazard(text: str) -> str:
|
||||
"""Compact a free-form NWS hazard sentence into the terse mesh idiom the
|
||||
SVR branch already uses, so no product type (SPS/WSW/FFW/FLW/else) carries a
|
||||
bloated line 3. Drops filler ('in excess of' -> '>'), collapses '45 mph' ->
|
||||
'45mph', rewrites wind-gust phrasings to 'Nmph gusts', and converts hail
|
||||
descriptors ('pea size hail') to numeric inches ('0.25" hail')."""
|
||||
if not text:
|
||||
return ""
|
||||
t = text.strip().rstrip(".")
|
||||
# Hail: '<descriptor> size hail' -> 'N" hail' (keep any leading connector).
|
||||
low = t.lower()
|
||||
for k in sorted(_HAIL_DESCRIPTORS, key=len, reverse=True):
|
||||
for variant in (f"{k} size hail", f"{k}-size hail", f"{k} sized hail"):
|
||||
idx = low.find(variant)
|
||||
if idx != -1:
|
||||
repl = _fmt_hail(k)
|
||||
if repl:
|
||||
t = t[:idx] + repl + t[idx + len(variant):]
|
||||
low = t.lower()
|
||||
break
|
||||
# Numeric hail: 'N inch hail' / 'N-inch hail' -> 'N" hail'.
|
||||
t = re.sub(r"(\d+(?:\.\d+)?)[\- ]inch(?:es)?\s+hail",
|
||||
lambda m: _fmt_hail(m.group(1)) or m.group(0), t,
|
||||
flags=re.IGNORECASE)
|
||||
# Wind-gust phrasings -> 'Nmph gusts'.
|
||||
t = re.sub(r"wind gusts?\s+(?:in excess of|up to|to|of|reaching|near|around)"
|
||||
r"\s+(\d+)\s*mph", r"\1mph gusts", t, flags=re.IGNORECASE)
|
||||
t = re.sub(r"winds?\s+gusting\s+(?:up\s+)?to\s+(\d+)\s*mph",
|
||||
r"\1mph gusts", t, flags=re.IGNORECASE)
|
||||
# Sustained winds -> 'Nmph winds'.
|
||||
t = re.sub(r"(?:damaging\s+)?winds?\s+(?:in excess of|up to|to|of)\s+"
|
||||
r"(\d+)\s*mph", r"\1mph winds", t, flags=re.IGNORECASE)
|
||||
# Remaining generic filler + spacing.
|
||||
t = re.sub(r"\bin excess of\b", ">", t, flags=re.IGNORECASE)
|
||||
t = re.sub(r"(\d+)\s*mph", r"\1mph", t, flags=re.IGNORECASE)
|
||||
t = re.sub(r"\s+", " ", t).strip()
|
||||
return t
|
||||
|
||||
|
||||
def _render(*, event_type, area_desc, geocoder_city, county, state,
|
||||
expires_epoch, lat, lon, now, prefix: str = "", d: dict = None) -> str:
|
||||
d = d or {}
|
||||
|
|
@ -443,9 +482,10 @@ def _render(*, event_type, area_desc, geocoder_city, county, state,
|
|||
line3 = f"{hazard} · {confirm}" if hazard else confirm
|
||||
elif same_code in ("FFW", "FLW"):
|
||||
hazard_text = desc.get("hazard") or ""
|
||||
# First sentence only
|
||||
# First sentence only, then tighten to the terse SVR idiom.
|
||||
if ". " in hazard_text:
|
||||
hazard_text = hazard_text.split(". ")[0]
|
||||
hazard_text = _tighten_hazard(hazard_text)
|
||||
# Infer flood cause from description
|
||||
desc_lower = (d.get("description") or "").lower()
|
||||
flood_cause = ""
|
||||
|
|
@ -459,10 +499,12 @@ def _render(*, event_type, area_desc, geocoder_city, county, state,
|
|||
cause_seg = f" · {flood_cause}" if flood_cause else ""
|
||||
line3 = f"{hazard_text}{cause_seg}" if hazard_text else flood_cause
|
||||
else:
|
||||
# SPS, WSW, etc.: first hazard sentence + certainty if Observed/Likely
|
||||
# SPS, WSW, etc.: first hazard sentence (tightened) + certainty if
|
||||
# Observed/Likely.
|
||||
hazard_text = desc.get("hazard") or ""
|
||||
if ". " in hazard_text:
|
||||
hazard_text = hazard_text.split(". ")[0]
|
||||
hazard_text = _tighten_hazard(hazard_text)
|
||||
cert_seg = ""
|
||||
if certainty in ("Observed", "Likely"):
|
||||
cert_seg = f" · {_collapse_certainty(certainty)}"
|
||||
|
|
@ -481,6 +523,15 @@ def _render(*, event_type, area_desc, geocoder_city, county, state,
|
|||
towns[-1] = re.sub(r"^and\s+", "", towns[-1], flags=re.IGNORECASE).strip()
|
||||
towns = [t for t in towns if t]
|
||||
|
||||
def _dedup(seq):
|
||||
"""Drop consecutive repeats (a short list can make first == middle),
|
||||
so we never render 'Buhl → Buhl → Shoshone'."""
|
||||
out = []
|
||||
for t in seq:
|
||||
if not out or out[-1] != t:
|
||||
out.append(t)
|
||||
return out
|
||||
|
||||
def _line4(locs: str) -> str:
|
||||
if motion and locs:
|
||||
return f"{motion} — {locs}"
|
||||
|
|
@ -490,32 +541,39 @@ def _render(*, event_type, area_desc, geocoder_city, county, state,
|
|||
|
||||
PACKET_LIMIT = budget_for("nws")
|
||||
|
||||
# Prefer the FULL town list; use it verbatim if the whole message fits.
|
||||
full_locs = ", ".join(towns)
|
||||
line4 = _line4(full_locs)
|
||||
msg = "\n".join(l for l in (line1, line2, line3, line4) if l)
|
||||
# Location representations from richest to poorest: full comma list ->
|
||||
# first→middle→last path sample -> first→last -> first-only -> none. The
|
||||
# WHOLE message is measured against the budget for each and the first form
|
||||
# that fits wins. Crucially the "— {locs}" segment is only ever attached
|
||||
# when the full message fits, so we can never emit a dangling "— …": if no
|
||||
# location form fits we fall through to motion-only, then (if even that
|
||||
# overflows) drop line 4 entirely.
|
||||
loc_options = [", ".join(towns)] # full comma list
|
||||
if len(towns) >= 3:
|
||||
loc_options.append(" → ".join(
|
||||
_dedup([towns[0], towns[len(towns) // 2], towns[-1]])))
|
||||
if len(towns) >= 2:
|
||||
loc_options.append(" → ".join(_dedup([towns[0], towns[-1]])))
|
||||
if towns:
|
||||
loc_options.append(towns[0])
|
||||
loc_options.append("") # motion only / empty
|
||||
|
||||
if len(msg) > PACKET_LIMIT:
|
||||
# Won't fit: collapse locations to a path sample that never loses the
|
||||
# endpoints -> soonest-impact -> midway -> farthest-along.
|
||||
if len(towns) >= 3:
|
||||
sampled = [towns[0], towns[len(towns) // 2], towns[-1]]
|
||||
elif len(towns) == 2:
|
||||
sampled = [towns[0], towns[-1]]
|
||||
else:
|
||||
sampled = list(towns)
|
||||
# De-dup consecutive repeats (a short list can make first == middle),
|
||||
# so we never render "Buhl → Buhl → Shoshone".
|
||||
deduped = []
|
||||
for t in sampled:
|
||||
if not deduped or deduped[-1] != t:
|
||||
deduped.append(t)
|
||||
path = " → ".join(deduped)
|
||||
line4 = _line4(path)
|
||||
msg = "\n".join(l for l in (line1, line2, line3, line4) if l)
|
||||
base_lines = [l for l in (line1, line2, line3) if l]
|
||||
msg = None
|
||||
for locs in loc_options:
|
||||
cand4 = _line4(locs)
|
||||
lines = base_lines + ([cand4] if cand4 else [])
|
||||
candidate = "\n".join(lines)
|
||||
if len(candidate) <= PACKET_LIMIT:
|
||||
msg = candidate
|
||||
break
|
||||
if msg is None:
|
||||
# Even motion-only line 4 overflows: drop line 4 entirely.
|
||||
msg = "\n".join(base_lines)
|
||||
|
||||
# Final hard-cap safety net for the pathological case (extremely long town
|
||||
# names overflow even the first->middle->last sample plus the other lines).
|
||||
# Final hard-cap safety net for the pathological case where lines 1-3 alone
|
||||
# overflow. Line 4 is already budget-fitted above, so this only ever trims
|
||||
# the leading lines — it can never manufacture a dangling "— …".
|
||||
return fit_to_budget(msg, PACKET_LIMIT)
|
||||
|
||||
|
||||
|
|
|
|||
|
|
@ -378,3 +378,166 @@ def test_svr_worst_case_fits_140(mem_db):
|
|||
assert "70mph winds" in rendered # tightened hazard (wind)
|
||||
assert '1.75" hail' in rendered # golf ball -> 1.75"
|
||||
assert "Buhl" in rendered # >=1 town present
|
||||
|
||||
|
||||
# ---- no dangling "— …" across ALL product types ----
|
||||
|
||||
|
||||
def _assert_no_dangling_separator(rendered: str):
|
||||
"""The L4 motion/locations line must never end in a stray separator:
|
||||
no trailing '—…', '— …', or a bare '—'. Either a real location list
|
||||
follows the em-dash, or the em-dash (and its locations) are absent."""
|
||||
for line in rendered.splitlines():
|
||||
stripped = line.rstrip()
|
||||
assert not stripped.endswith("—…"), f"dangling '—…': {line!r}"
|
||||
assert not stripped.endswith("— …"), f"dangling '— …': {line!r}"
|
||||
assert not stripped.endswith("—"), f"bare trailing '—': {line!r}"
|
||||
# And the "— …" fragment must not appear mid-line either.
|
||||
assert "—…" not in stripped, f"'—…' fragment: {line!r}"
|
||||
assert "— …" not in stripped, f"'— …' fragment: {line!r}"
|
||||
|
||||
|
||||
def test_sps_worst_case_tightened_and_no_dangling(mem_db):
|
||||
"""The reported live-log bug: a Special Weather Statement (SPS) with wind
|
||||
gusts + motion + a long town list previously collapsed L4 to
|
||||
'Moving SW 24 mph —…' (all towns lost, dangling separator). After the fix:
|
||||
hazard is tightened, output fits 140, and L4 is either
|
||||
'Moving … — <towns>' or 'Moving …' — never a trailing '—…'."""
|
||||
long_locations = (
|
||||
"Twin Falls, Kimberly, Filer, Buhl, Hansen, Murtaugh, Hollister, "
|
||||
"Eden, Hazelton, and Rogerson"
|
||||
)
|
||||
description = (
|
||||
"HAZARD...Wind gusts in excess of 45 mph and pea size hail.\n\n"
|
||||
"SOURCE...Radar indicated.\n\n"
|
||||
f"Locations impacted include...{long_locations}"
|
||||
)
|
||||
d = {
|
||||
"eventCode": {"SAME": ["SPS"]},
|
||||
"certainty": "Observed",
|
||||
"parameters": {
|
||||
# 225 DEG 21 KT -> "Moving SW 24 mph"
|
||||
"eventMotionDescription": ["2200000T225DEG...21KT 42.5,-114.5"],
|
||||
},
|
||||
"description": description,
|
||||
}
|
||||
rendered = _render(
|
||||
event_type="Special Weather Statement", area_desc="Twin Falls County",
|
||||
geocoder_city=None, county="Twin Falls", state="ID",
|
||||
expires_epoch=1_751_400_000, lat=42.5, lon=-114.46,
|
||||
now=1_751_400_000, d=d,
|
||||
)
|
||||
assert len(rendered) <= 140, f"{len(rendered)} chars:\n{rendered!r}"
|
||||
assert "Special Weather Statement" in rendered
|
||||
# (a) hazard tightened: "Wind gusts in excess of 45 mph" -> "45mph gusts",
|
||||
# "pea size hail" -> '0.25" hail'; filler dropped.
|
||||
assert "45mph gusts" in rendered, f"wind not tightened:\n{rendered!r}"
|
||||
assert '0.25" hail' in rendered, f"hail not numeric:\n{rendered!r}"
|
||||
assert "in excess of" not in rendered, "filler 'in excess of' survived"
|
||||
assert "· observed" in rendered, "certainty not collapsed"
|
||||
# (b) NEVER a dangling separator.
|
||||
_assert_no_dangling_separator(rendered)
|
||||
# (c) the motion line, when present, either carries a town or stands alone.
|
||||
last = rendered.splitlines()[-1]
|
||||
if last.startswith("Moving"):
|
||||
assert last == "Moving SW 24 mph" or " — " in last, (
|
||||
f"L4 neither motion-only nor motion+towns:\n{last!r}")
|
||||
if " — " in last:
|
||||
# A real town must follow the em-dash.
|
||||
tail = last.split(" — ", 1)[1].strip()
|
||||
assert tail and tail != "…", f"empty tail after em-dash:\n{last!r}"
|
||||
|
||||
|
||||
def test_wsw_hazard_tightened_and_no_dangling(mem_db):
|
||||
"""Winter Weather product (WSW SAME code): wind-gust hazard is tightened and
|
||||
no dangling '—…' can appear."""
|
||||
long_locations = (
|
||||
"Sun Valley, Ketchum, Hailey, Bellevue, Carey, Picabo, Fairfield, "
|
||||
"and Gooding"
|
||||
)
|
||||
description = (
|
||||
"HAZARD...Wind gusts up to 45 mph and heavy snow.\n\n"
|
||||
f"Locations impacted include...{long_locations}"
|
||||
)
|
||||
d = {
|
||||
"eventCode": {"SAME": ["WSW"]},
|
||||
"certainty": "Observed",
|
||||
"parameters": {
|
||||
"eventMotionDescription": ["2200000T225DEG...21KT 42.5,-114.5"],
|
||||
},
|
||||
"description": description,
|
||||
}
|
||||
rendered = _render(
|
||||
event_type="Winter Weather Advisory", area_desc="Blaine County",
|
||||
geocoder_city=None, county="Blaine", state="ID",
|
||||
expires_epoch=1_751_400_000, lat=43.5, lon=-114.3,
|
||||
now=1_751_400_000, d=d,
|
||||
)
|
||||
assert len(rendered) <= 140, f"{len(rendered)} chars:\n{rendered!r}"
|
||||
assert "45mph gusts" in rendered, f"WSW wind not tightened:\n{rendered!r}"
|
||||
assert "heavy snow" in rendered
|
||||
assert "up to" not in rendered, "filler 'up to' survived"
|
||||
_assert_no_dangling_separator(rendered)
|
||||
|
||||
|
||||
def test_sps_pathological_towns_degrade_to_motion_only(mem_db):
|
||||
"""When even a single sampled town cannot fit the remaining budget, L4 must
|
||||
degrade to motion-only ('Moving …') with NO trailing separator — never
|
||||
'Moving … —…'."""
|
||||
# One absurdly long town name that cannot coexist with the em-dash + motion
|
||||
# in the leftover budget.
|
||||
long_town = "Averyverylongimpossibletownnamethatwillnotfitthebudgetatall" * 2
|
||||
description = (
|
||||
"HAZARD...Wind gusts in excess of 45 mph.\n\n"
|
||||
f"Locations impacted include...{long_town}"
|
||||
)
|
||||
d = {
|
||||
"eventCode": {"SAME": ["SPS"]},
|
||||
"certainty": "Observed",
|
||||
"parameters": {
|
||||
"eventMotionDescription": ["2200000T225DEG...21KT 42.5,-114.5"],
|
||||
},
|
||||
"description": description,
|
||||
}
|
||||
rendered = _render(
|
||||
event_type="Special Weather Statement", area_desc="Twin Falls County",
|
||||
geocoder_city=None, county="Twin Falls", state="ID",
|
||||
expires_epoch=1_751_400_000, lat=42.5, lon=-114.46,
|
||||
now=1_751_400_000, d=d,
|
||||
)
|
||||
assert len(rendered) <= 140
|
||||
_assert_no_dangling_separator(rendered)
|
||||
last = rendered.splitlines()[-1]
|
||||
# The town can't fit, so L4 (if present) is bare motion.
|
||||
if last.startswith("Moving"):
|
||||
assert " — " not in last, f"expected motion-only, got:\n{last!r}"
|
||||
|
||||
|
||||
def test_svr_no_dangling_separator(mem_db):
|
||||
"""Re-verify SVR (the branch tightened earlier) still never dangles."""
|
||||
long_locations = (
|
||||
"Buhl, Eden, Hazelton, Murtaugh, Richfield, Dietrich, Gooding, "
|
||||
"Hagerman, Wendell, Jerome, Kimberly, Hansen, Filer, and Shoshone"
|
||||
)
|
||||
description = (
|
||||
"HAZARD...Damaging winds to 70 mph and golf ball size hail.\n\n"
|
||||
f"Locations impacted include...{long_locations}"
|
||||
)
|
||||
d = {
|
||||
"eventCode": {"SAME": ["SVR"]},
|
||||
"certainty": "Observed",
|
||||
"parameters": {
|
||||
"maxWindGust": ["70 MPH"], "maxHailSize": ["1.75"],
|
||||
"eventMotionDescription": ["2200000T254DEG...35KT 42.5,-114.5"],
|
||||
},
|
||||
"description": description,
|
||||
}
|
||||
rendered = _render(
|
||||
event_type="Severe Thunderstorm Warning", area_desc="Twin Falls County",
|
||||
geocoder_city=None, county="Twin Falls", state="ID",
|
||||
expires_epoch=1_751_400_000, lat=42.5, lon=-114.46,
|
||||
now=1_751_400_000, d=d,
|
||||
)
|
||||
assert len(rendered) <= 140
|
||||
assert "70mph winds" in rendered
|
||||
_assert_no_dangling_separator(rendered)
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue