Commit Graph
11 Commits
Author SHA1 Message Date
Sameh Khamis df5ad4ca4a Pair TLE lines through the shared decoder and add the changelog fragment 2026-10-17 13:46:00 -07:00
Sameh Khamis 155f5165d5 Merge branch 'main' into fix/tle-catalog-matching 2026-10-17 13:45:09 -07:00
Sameh Khamis 7ec84ac316 Decode catalog numbers strictly and add the changelog fragment
Route tleCatalogNumber() and noradIdFromSatnum() through one decoder that
accepts only plain digits or a valid Alpha-5 field. parseInt and Number()
read '2554A', '255 4', '-1234', '0x1F0', '1.5' and blank as numbers; they
now return null/NaN. Digit satnums of any length still decode, so OMM
satrecs numbered past 99999 keep working.

Add a layer test that several Alpha-5 satellites survive ingestion dedupe,
the changelog fragment for #767, and note in CURRENT-STATE that this is
Alpha-5 TLE compatibility and malformed fields are rejected.
2026-10-17 11:40:30 -07:00
VG31OP 40442ab854 style: fix whitespace formatting in tle.js and space.test.mjs 2026-10-17 17:01:13 +05:30
VG31OP b48f44f6e8 Merge remote-tracking branch 'origin/main' into fix/tle-catalog-matching 2026-10-16 22:44:46 +05:30
Honorius Neogy ac62542641 fix(satellites): decode Alpha-5 catalog numbers in the shared TLE module
Catalog numbers above 99999 arrive in a TLE as Alpha-5: a letter A-Z,
skipping I and O, stands for 10-33 in the ten-thousands place, so 100000
is A0000. satellite.js keeps the field as the string satrec.satnum, and
the three catalog builders converted it with Number(), which gives NaN;
the Set and Map used to deduplicate and store treat NaN as equal to NaN,
so the first such satellite was stored under NaN and every later one was
dropped as a duplicate (#751). The satellite tools read the same field
with parseInt and reported no number, so such a satellite could not be
found by it.

The decoder lives in src/sources/tle.js, beside parseTleText, which the
layer and the tools both import already: noradIdFromSatnum() for the
three builders, and tleCatalogNumber() decodes through the same function.
Every other field reads as it did before: Number() in the layer,
parseInt or null in the tools.
2026-10-15 18:15:36 -04:00
VG31OP 003b92f64a fix(satellites): reject TLE pairs with mismatched catalog numbers (#906, #936) 2026-10-13 16:57:55 +05:30
Huanyi Xie 1c3748e3bd fix(satellites): recover TLE parsing after incomplete sets 2026-10-05 13:02:20 +03:00
Sameh Khamis e779432925 fix: report stale and partial data in more tool answers
A sweep for the same pattern found more answers presented as current and
complete: launches and satellite orbits served from a stale proxy cache, a
regional aircraft fallback read as worldwide, license plate cameras trimmed
or partly loaded, and a radio directory of popular stations read as every
station. The launch and satellite sources now report the proxies' stale
markers, and each answer says what it covers.
2026-10-02 15:21:36 -07:00
Sameh Khamis ec1adbdc19 fix: tighten tool answers found in local testing
Unknown aircraft tracks read as empty instead of a server error, aircraft
info names the aircraft and flight, the HUD caption sends the app's caption
context, and satellite pass, satellites overhead and nearby places take a
place name or point like the other tools.
2026-10-01 14:01:25 -07:00
Sameh Khamis 22c3fe40b0 feat: add aviation and satellite tools
Add aircraft_in_area, find_aircraft, get_aircraft_track and
get_aircraft_info over the live aircraft sources, with an optional
military feed, and next_satellite_pass and satellites_overhead over the
satellite catalog and the existing pass predictor.

Source failures from the live contract now become tool errors with the
matching code, so a rate limit reaches clients as retry_later with a
wait time. Tool schemas may use pattern.
2026-10-01 12:05:50 -07:00