Repository navigation
Conversation
mrosseel
force-pushed
the
deepchart
branch
6 times, most recently
from
June 24, 2026 19:57
b31666e to
19717a2
Compare
mrosseel
force-pushed
the
deepchart
branch
11 times, most recently
from
July 11, 2026 22:26
eb23e35 to
a4c44c2
Compare
The colour term matches the stars to the sensor passband, so the sky is measured in that passband too. A bare sensor sees NIR sky emission (airglow, LED/sodium LP beyond 700nm) that a V-band SQM meter does not, so its sky reads genuinely brighter. sqm_band_offset converts the final SQM back to the meter scale: imx462/imx290 +0.43, hq +0.07, imx296 0.0 (unmeasured). Earlier a similar constant was removed because it papered over 0.6 mag of variable systematics. With the black level, colour term, wing correction and robust mzero in place the residual is now genuinely constant (night-to-night spread 0.06 mag across focus states and star fields), so a per-sensor value is justified. Values remain provisional (3 sweeps imx462, 2 trusted sweeps hq); refine with a multi-night side-by-side campaign against a reference meter aimed at the camera field. Sweep validation with the offset applied: imx462 -0.03/+0.03/+0.01 vs meter; hq +0.03/-0.03 on trusted nights. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VaxurSrCzxrsfJrSjfA5bo
…sted Extract the fake GPS fix into apply_test_mode_gps() and call it during startup when test_mode is saved in config, so a reboot in test mode comes back with location and datetime set, not just the debug camera.
The colour term matches the stars to the sensor passband, so the sky is measured in that passband too. A bare sensor sees NIR sky emission (airglow, LED/sodium LP beyond 700nm) that a V-band SQM meter does not, so its sky reads genuinely brighter. sqm_band_offset converts the final SQM back to the meter scale: imx462/imx290 +0.43, hq +0.07, imx296 0.0 (unmeasured). Provisional values; refine with a side-by-side campaign. Sweep validation with the offset applied: imx462 -0.03/+0.03/+0.01 vs meter; hq +0.03/-0.03 on trusted nights. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VaxurSrCzxrsfJrSjfA5bo
Test mode was split across two independently-toggled flags: a persisted test_mode config/shared-state flag and a camera-process-local debug flag flipped by a "debug" queue command. After a reboot with test mode persisted ON the camera always came back with its flag OFF, so the menu showed Test Mode ON while the solver got real (unsolvable indoor) frames — and further toggles could never re-sync the two. The camera loop now reads shared_state.test_mode() directly each iteration; the "debug" camera command and its senders are removed. Startup with persisted test mode also applies the fake GPS fix (extracted to apply_test_mode_gps), so a reboot in test mode comes back fully in test mode. The console 0-key easter egg now routes through the same ui_queue toggle. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FrqHHkXpVHhNWTKZEb63yH
…#109) Opening the Messier list filtered every catalog, WDS included (131,303 objects). With an altitude criterion the filter took 2.4 s on a CM4, one object at a time in Python. - CatalogFilter.verdicts evaluates each active criterion for all objects at once with numpy. FastAltAz.radec_to_alt_array is the numpy form of the altitude calculation. WDS now takes 0.5 s on the CM4. - Catalogs.filter_catalogs takes the catalogs to filter. An object list filters only the catalogs it shows, so the Messier list filters 110 objects. The chart filters the selected catalogs before it builds its nearby index. - An object list rebuilds when the filter's dirty_time differs from the one it was built with, so it does not depend on which catalog was filtered last. Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Each screen kept its own frame counter. After a screen change, the new screen's first value covered only part of a second, and a screen you went back to counted from its old sample time. The title bar showed a low FPS for about one second after each change, also when the real rate did not drop. FrameRate is one counter on UIModule. The menu manager counts each frame it sends to the display, and the rate divides by the real window length. Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
#114) The early boot splash set the SSD1333 to contrast 0xFF and master current 0x0F: a drive product of 4080. That is 2.3 times the blooming cap in DisplaySSD1333 (1792), so bright pixels smeared to the left. The app splash then showed the same image clean, at set_brightness(125). The boot splash now sets contrast 142 and master 1, the registers of set_brightness(125), so the image also keeps its brightness when the app takes over. Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
#111) While catalogs loaded, the UI ran at 2 FPS on the CM4. The UI thread waits for the shared-state process many times per frame. After each wait it needs the GIL back, and the catalog loader thread kept it for up to the 5 ms switch interval, so each call cost up to 5 ms. - The UI process sets the GIL switch interval to 1 ms, after the worker processes start. - A frame makes about 7 shared-state calls in place of 14. The popup time, the FPS flag and the hardware capabilities are kept in the UI process. The title bar reads SQM, constellation and battery once a second. The power state is read once per frame. The UI state and the frame go to shared state only when they change, through UIModule.publish_screen. Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…mns built in CI (#113) * perf(ui): keep the UI responsive while a thread in its process is busy While catalogs loaded, the UI ran at 2 FPS on the CM4. The UI thread waits for the shared-state process many times per frame. After each wait it needs the GIL back, and the catalog loader thread kept it for up to the 5 ms switch interval, so each call cost up to 5 ms. - The UI process sets the GIL switch interval to 1 ms, after the worker processes start. - A frame makes about 7 shared-state calls in place of 14. The popup time, the FPS flag and the hardware capabilities are kept in the UI process. The title bar reads SQM, constellation and battery once a second. The power state is read once per frame. The UI state and the frame go to shared state only when they change, through UIModule.publish_screen. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * docs: UI performance plan (state snapshot, catalog arrays) * perf(state): read shared state from a shared-memory snapshot The UI thread made a round trip to the shared-state manager for each value it read, many times per frame. During each wait it released the GIL, and a busy thread in the UI process could then keep the GIL for up to the switch interval. The manager process now also writes the most-read values into one shared-memory block, one slot per value with its own lock (PiFinder/state_snapshot.py). The main loop reads a StateSnapshot once per frame into UIModule.snapshot, and the screens read from it. A read is a memory copy: no round trip, and the GIL is not released. The setters and their rules do not change. - StateSnapshot has the getter names of SharedStateObj, so read-only helpers take either one. CurrentSnapshot gives the catalog filter and Nearby the snapshot of the current frame. - A heartbeat slot makes a reader raise SharedStateLost when the manager stops, as the proxy does. - PowerManager keeps the power state it writes. - In a headless run, the UI main thread makes no manager read in a normal frame (was about 7). Design: ui-performance-plan.md, step 1. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * perf(catalogs): catalogs as numpy columns, built in CI The UI process kept one CompositeObject per catalog listing: 151,170 objects, about 276 MB, plus a 46 MB T9 cache. A loader thread built the WDS objects after start and took the GIL from the UI thread for about 17 s. Filters, search, All Filtered and the Nearby index walked every object in Python. - catalog_arrays.py builds one directory per DB catalog: numpy columns and UTF-8 text blobs, opened with mmap. A Nix derivation (nixos/pkgs/catalog-arrays.nix) builds them in CI from the DB and the builder files only. Without Nix, the app builds them into the cache directory once. - ArrayCatalog makes a CompositeObject only for a row that code reads, and keeps the last 5000. logged, the filter verdict and the observing-list descriptions live next to the columns. - ObjectSequence is a read-only list over catalog rows. Filter, sort, Nearby de-duplication, search and the list cursor work on columns. - Planets, comets and observing-list objects stay plain Catalogs. - Gone: the loader threads, priority and deferred catalogs, the pickle cache (and its files), the T9 cache, the Names class and the "Catalogs Fully Loaded" message. On a CM4: filter WDS 150-200 ms (was 2.37 s), all catalogs 170 ms, Nearby index 220 ms, open 0.8-0.9 s with no thread after start. Private RAM of the UI process in a headless PC run: 152 MB (was 542 MB). Every listing that ArrayCatalog makes is tested against the rules of the per-object builder, for the whole DB. Design: ui-performance-plan.md, step 2; ADR 0043. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * fix(ui): show the sorting popup only when the user changes the sort Opening a catalog list called sort(), which always showed a "Sorting by" popup for 0.1 s. The popup went to the display at once on the empty new screen, and while it showed, update_screen skipped the frames of the 0.1 s menu slide. So the list appeared after a popup, with no slide. sort() shows the popup only with announce=True, which the marking-menu sort choices pass. Opening and refreshing a list show no popup, and the slide is visible. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* feat(display): demo display for instruction videos Add the pg_demo and pg_demo_176 displays. The window shows a photo of the PiFinder with the live UI in its screen area. Each key that the main loop reads makes its button glow. A click on a button in the photo presses that key. A right click is a press and hold, and Ctrl + click is a SQUARE chord. --record FILE writes the window to an MP4 file through ffmpeg at 30 fps. --record-audio adds the default microphone. python/scripts/demo.sh starts it all with one command, and scripts/demo_tour.pfs is an example key script for a video you can record again. This replaces the separate overlay program, which read one fixed keyboard with evdev and needed OBS to put the emulator window on the photo. See docs/adr/0042-demo-display-in-process.md. The dev shells get ffmpeg. The Pi image has no pygame, so it does not change. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * fix(nix): use the new names for the X11 libraries nixpkgs 26.05 deprecates the xorg package set. The pygame override used pkgs.xorg.libX11 and five other old names, so each dev shell load showed six evaluation warnings. The new names give the same store paths, so the build output does not change. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * fix(nix): keep ffmpeg out of the dev shells ffmpeg adds about 1 GB to every dev shell, and only a demo recording needs it. The recorder error now names nix shell nixpkgs#ffmpeg-headless, and the dev guide and ADR 0042 say the same. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * feat(demo): rename the start script to record_demo.sh and show its options The script now shows its options and the keys before PiFinder starts, and -h shows only that text. A relative --record path is now relative to the folder where you start the script, not to python/. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* fix(catalogs): precess Sharpless declinations to J2000
load_sharpless computed j_dec_deg from b1950_to_j2000() and then passed the
unconverted dec_deg to NewCatalogObject, so only right ascension was
precessed. All 313 Sharpless objects sat 0.26 to 0.37 degrees off in
declination, which is exactly the 1950-to-2000 shift at those coordinates.
Found by cross-checking against the Perek-Kohoutek catalogue, which lists
several of the same nebulae with independent J2000 positions.
* fix(catalogs): correct two Abell cross-identification errors
abell.tsv carried the ngc6742 alias on row 51. NGC 6742 matches row 50 to
0.1 arcmin and is 66 degrees from row 51, and the Strasbourg-ESO catalogue
independently gives A 50, not A 51, as NGC 6742. The misplaced alias bound
the Abell 51 listing onto NGC 6742's sky object.
Row 47 had lost its declination sign. Abell 47 is at -00 13 51 per SIMBAD
and Kohoutek 2001. Its constellation, Serpens, spans the celestial equator,
so nothing else caught it.
The two remaining aliases were audited positionally: IC 972, NGC 7076 and
IC 1454 all agree within 0.12 arcmin.
* feat(catalogs): add the Perek-Kohoutek planetary nebula catalog
Adds catalog code PK: the 1510 galactic planetary nebulae of Kohoutek's
2001 revision of Perek-Kohoutek 1967, from VizieR IV/24. Positions come
from SIMBAD or IV/24 table4, sizes and cross-identifications from V/84.
Sources are vendored under astro_data/perek_kohoutek/ with PROVENANCE.md
and a fetch script; decisions are in docs/adr/0024.
No precession runs here. Every position used is already J2000, and each
refined position must agree with the catalogue's own coarse position
within 5 arcmin, which rejects a four-degree typo in IV/24 table4.
The sequence is the printed catalogue's 1-1510 running number, so "PK 743"
is a list position, not a designation; the real designation is stored as a
name in both spellings. No PK source carries a magnitude, so entries are
built with an empty one: the 211 that resolve to an existing NGC, IC,
Messier, Abell or Sharpless object inherit its magnitude, and the rest fall
out of any magnitude filter via UNKNOWN_MAG.
Alias matching had to be fixed first. ObjectFinder resolved aliases through
ui_utils.normalize(), which strips spaces and hyphens, so any compound
numeric part collapsed into a wrong sequence number. Feeding it this
catalogue's name column produced 147 matches, 145 of them false - 188 of
those names are Minkowski planetary nebulae ("M 1-92"), not Messier
objects, and "H 3-29" is Haro, not Herschel 329. catalog_import_utils now
carries CATALOG_CODE_ALIASES and parse_designation(), which require an
explicitly allowlisted prefix and a bare integer sequence, so a compound
designation never reaches hyphen-stripping.
That fix also makes an existing intent work: post_processing lists "Cr 42"
among M45's aka names so the Pleiades resolve to one object, but Cr had
never been mapped to Col, so the database shipped two Pleiades. It also
removes two false links in SAC Multistars.
Verified against the published 1510-row list transcribed on astronomy.com
and in a widely circulated spreadsheet: all 1510 PK designations and all
1510 other-designations match exactly. That check also found the
spreadsheet drops a declination sign at row 551.
* fix(catalogs): join a split NGC cross-identification in the PK import
V/84 writes the cross-identification of PK 078+18.1 as "NGC,6742". The
import split the field on commas, so NGC 6742 got the names "NGC" and
"6742". _split_idents joins a token without a digit with a next token
of only digits; the Blanco names ("Bl G") stay as they are.
The committed DB has the two names removed. A full import cannot be
used here: it is not reproducible (TAAS 200 links and WDS name order
change from run to run), so it would add unrelated changes. A full
import with this code gives the same PK names.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* fix(catalogs): PK review fixes, and a reproducible catalog import
- pk_loader: the PNG column holds "poss.", "rej." or "--" when a nebula
has no Strasbourg-ESO designation. The loader checked for "possible"
and "rejected", so 185 entries got names like "PN Gposs.". Only a
designation becomes a PN G name now; the description states the
status, and f_PK '*' (possible PN per Kohoutek) is stated too.
- pk_loader: SIMBAD aliases only when SIMBAD types the object as a PN. A
cross-link is accepted only within 30 arcmin of the entry (the largest
correct link is 17 arcmin). No current link is rejected.
- pk_loader: ra_to_deg and dms_to_dec from calc_utils replace its own
copies.
- parse_designation: no global "a" and "b" (Aitken and van den Bos in
double-star sources); a loader passes extra_aliases instead. The
allowlist holds every other PiFinder catalog code, so "Arp 244" links.
- load_taas200 and load_wds keep the source order of their names. Two
imports with different hash seeds now give identical tables. Ta2 183
joins NGC 6960 and Ta2 52 joins NGC 2237.
- post_processing: M45 keeps the curated Pleiades values on the object
it shares with Col 42.
- obslist: the name index keys on the normalized name and its digit
groups, so "M 2-9" no longer resolves to M 29.
- Docs: ADR 0045 describes the fixes; no em dashes in prose; no history
in the test docstring.
The DB is a full rebuild with this code. Against the previous DB: 186
PK names removed (185 "PN G" status names, and one SIMBAD alias of a
non-PN), PK descriptions updated, M45 values, and links changed only
for Ta2 52 and Ta2 183.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
- catalogs.py: the background loader is gone in nixos (catalog arrays, #113), so its muted progress log goes with it. - object_details.py: deepchart's object_images imports plus ObjectSequence; numpy is imported once at the top, not in functions. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
The Gaia chart generator is a singleton that keeps the state object of its first caller, and main.py created it with the shared-state proxy. So every chart and every cache key read SQM (and location) through a manager round trip. main.py and object details now pass CurrentSnapshot(), the view of the current frame's state snapshot; the generator only reads. The log screen reads its SQM note from self.snapshot as well. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
) 2.6.1 to 2.6.3 read migration_url and migration_sha256_url from the update manifest. With the 7x square they then install a btrfs tarball with the old ext4 migration code, and the card does not boot. The generator now writes migration_v2_url and migration_v2_sha256_url. Only the new migration code (brickbots#657) reads these keys. Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…rver (#117) The brickbots DNS now points cache.pifinder.io and deltas.pifinder.io at the same server as the .eu names. The devices, the CI and the docs use the .io names from this change on. The server keeps the .eu names as aliases. Devices that run an older build still use them. Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
On rev 4, GPIO14 is the LTC2954 power-off kill line (ADR 0007). The config.txt sets enable_uart=1, which makes GPIO14 the mini UART TXD. The first byte that U-Boot prints cuts the power, so a migrated CM4 switches itself off at each power-on. - config.txt: a [cm4] section after [all] sets gpio=14=ip,pu. The firmware sets the pin back to an input with a pull-up before U-Boot starts. U-Boot needs enable_uart=1, and its serial drivers check only GPIO15, so it keeps its console. Brickbots tested this on a rev 4 unit: a cold boot stays on, reboot comes back, poweroff cuts the power. - CM4 device tree: the gpio-poweroff overlay disables uart1. The kernel tree marked it okay with stdout-path on it, so the kernel registered ttyS0 on GPIO14 as its default console. - SD images: nixos/sd-image.nix now holds the settings of nixpkgs' sd-image-aarch64.nix without its console=ttyS0 and console=ttyAMA0 kernel parameters. A list option merges, so a later module cannot remove them. The image derivation diff shows only that change. Already migrated rev 4 cards keep their old config.txt, because an upgrade does not write the FAT partition. Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…igration guide (#119) A release builds a migration tarball only when new_migration_tarball is set. The tarball installs the newest release at first boot, so one tarball serves many releases. Without a new tarball, update_manifest.py --migration-from-manifest gives the entry the migration_v2 keys of the newest release that has them. With no such release it stops with an error. nixos/RELEASE.md now covers where each part is, the release rules, the migration steps, the fixed contract with Pi OS 2.6.4, when a new tarball is needed, and the known faults. ADR 0039 names migration_v2_url and the shared tarball. Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* chore(nixos): kernel 6.18.60 (Raspberry Pi stable_20261103), update nixpkgs and nixos-hardware The kernel tag and hash are fixed in nixos/pkgs/pifinder-kernel.nix, so the kernel changes only when that file changes. nixos-hardware now follows our nixpkgs. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> * fix(nixos): btrfs checksum modules in the initrd for kernel 6.18 The forced initrd module list also removes the modules that the NixOS btrfs module adds. Kernel 6.12 had crc32c built in. In 6.18 it is the module crc32c-cryptoapi, so stage 1 could not mount the btrfs root. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
* docs(nixos): kernel update steps and the initrd module fault * docs(nixos): use git commands in the release guide Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…122) nixos-hardware now sets sdImage.populateFirmwareCommands with mkForce. Our value had the same priority, so both scripts ran. Its read-only copies made our copy fail with Permission denied. mkOverride 40 keeps only firmwareCommands. Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Developers, pre-commit and CI now run the same checks in the same flake dev shell, built from python/uv.lock. - python/scripts/check.sh runs each stage and enters the dev shell when needed. pytest runs with an empty temporary HOME, as on a CI runner. - ruff 0.4.8 is in the uv dev group. uv uses the Nix environment and makes no .venv. - nixos/libcamera-overlay.nix is used by the system and the dev shells, so both use the same libcamera. - .github/actions/setup-nix is the one Nix and cache setup for all jobs. - .github/scripts/push_dev_shell.sh pushes a dev shell's inputDerivation to the cache: nox (aarch64) and a new x86 job on green pushes, PR builds and releases. - nox and the web tests run on ubuntu-24.04-arm. The web dev shell adds chromium and chromedriver. - Tests: Blinka runs as a generic PC, every test gets sys_utils_fake, test_sys_utils.py tests the NixOS WiFi list and set_host_name, and the UI harness skips time.sleep in the main thread only. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Brings in 24 commits of brickbots main: the catalog import fixes (brickbots#671 to brickbots#694), the Named Stars catalog (brickbots#695), the names view of the object list (brickbots#696), the Pi OS side of the migration (brickbots#657), the release workflow (brickbots#658), the OLED SPI clock (brickbots#701) and CI changes. Conflicts and their resolution: - Pi OS migration files (migration_gate.json, migration_progress, migration_progress.c, nixos_migration_calc.py): nixos keeps them deleted. They belong to the Pi OS line. - noxfile.py, requirements_dev.txt: nixos keeps them deleted. The flake dev shell and uv2nix replace them. - ui/software.py, tests/test_software.py, sys_utils.py, main.py: nixos keeps its own software screen. Its keep_awake and on_battery are the same feature as upstream's. The migration UI stays out of nixos. - catalog_imports: upstream's code. The parser moved to designations.py, so the local copy of parse_designation goes. The Perek-Kohoutek loader stays and runs before the Named Stars loader. - catalogs.py: upstream's NAME_MODE_CATALOGS. The T9 tables stay in catalog_arrays.py. - object_details.py: upstream's listing index (brickbots#678), with a search on the listing_key column for an ObjectSequence. - obslist.py and its tests: the name index of nixos (resolvers) with upstream's ambiguous designation check. - tests/test_ui_modules.py: upstream's comet retry stub, the catalog build of nixos (no background loader). - docs/ax/catalog.md, CONTEXT.md: text of both sides. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The three commits of brickbots release that main lacks: the back piece with edge inserts, the BOM warning about the Adafruit 2472 BNO055 board (brickbots#641) and the 2.6.2 pause of version.txt. The tree of main has the same content already. version.txt stays 2.6.3. Fixes for the merge of upstream main: - object_details.py: the listing index in the format of ruff 0.4.8. - star_identity.py: load_hipparcos copies the position columns. With copy-on-write, pandas gives a read-only array, and the fill of the unsolved positions failed. - test_obslist_designations.py: the fake catalogs give iter_names, as the name index of nixos reads it. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Brings the merge of brickbots main into nixos into deepchart. Conflicts and their resolution: - .github/workflows/nox.yml, web-integration-tests.yml: the aarch64 runner and check.sh of deepchart. The ubuntu-24.04 pin of upstream is for setup-python, which these workflows do not use. - ui/object_details.py: the image imports of deepchart, with listing_keys for the listing index of upstream (brickbots#678). One numpy import. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…fixes The committed database was built before the catalog import changes of upstream main. It now holds the Named Stars catalog (578 stars) and the import fixes (brickbots#681, brickbots#683, brickbots#685, brickbots#691, brickbots#692, brickbots#694, brickbots#695). EGC has 102 objects (was 91) and SaM 2157 (was 2162). IC 1805 is now one object with Sh2 190 and Lyn 77, of type cluster with nebula. The catalog data test has the new values. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Harris and Lynga loaders parsed their float columns as float32, and arp.sqlite stores its magnitudes as float32 values, so the database held values such as 6.46999979019165 for M 2. The Harris and Lynga loaders now parse float64, and the Arp loader rounds to two decimals. The database is rebuilt. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Add a gaia achart view to all objects, down to mag 17 but with an auto limiting magnitude taking into account SQM and equipment. This auto limit is overridable.
TODO: