Simutrans TimetablesCZ · fork 122.0 · pak128 VZ
mirror/pipeline/odis-regional-opavsko/README.mdView on GitHub ↗

ODIS Opavsko, Vítkovsko, Hlučínsko + MHD Opava (regional and town buses, trolleybuses; vypravenost sheet)

Output: ODIS Opavsko Vitkovsko Hlucinsko.html in the project root, published at the URL in config.json → artifact_url (republish to the same URL). Built on 21 Oct 2026 after the owner asked for the next ODIS sheets ("Yes, do those next sheets") and then: "Do include trolleybuses - consider a different method but do include them".

This folder holds only the sheet's config, data and authoring scripts. The pipeline scripts are the shared ones in ../odis-regional-hko/scripts/, run with --base odis-regional-opavsko. The method is ../bus-schedule-method.md (v3); the Opava town network (buses and trolleybuses) takes the same method as a small town.

Scope

Section Licences Operator Fleet Pak
Opavsko 902xxx Z-Group bus, pr. Opava 22- VZ-ODIS-bus
Vítkovsko 903xxx Z-Group bus, pr. Opava 23- VZ-ODIS-bus
Hlučínsko 904xxx Transdev Slezsko pr. Ostrava (some 24- buses registered under Transdev Morava) 24- VZ-ODIS-bus
MHD Opava, trolleybuses 905201–905204 (mhd-opava-trol, prefix 90520) Městský dopravní podnik Opava 3xx VZ-DPOpava-trolleybus
MHD Opava, buses 905xxx MDPO 1xx, 40x VZ-DPOpava-bus
  • 70 ODIS numbers: 59 modelled as 73 Simutrans lines, 126 vehicles (116 seen on the 20 Oct live map, 10 designed). Was 125 vehicles before the 25 Oct hand check of 281, 282, 290 and 292.
  • After the 25 Oct check of the rest of the sheet (below): 72 Simutrans lines (285 Dobroslavice folded into 285), and the schedule (05_schedule) needs 133 buses, 7 more than 126; the roster is not rebuilt yet.
  • Not modelled (config.unmodelled): 214, 256, 275 (sparse / shift, stops on other lines), 283 (two trips), 228 and 229 (night lines).
  • Lines shared with other sheets (the sheet with most trips models them): 615, 620, 691, 692, 693 are on the Novojičínsko/Bílovecko sheet; this sheet models 260 and 920 (also Bruntálsko/Rýmařovsko licences), 261, 263, 270, 285, 678 (also Novojičínsko/Bílovecko licences) and 555 (904555; the HKO sheet lists it).

Sources and what is special here

  • Trolleybuses: 905201–905204 are not in the bus JDF. jdf.extra_zips adds CIS JŘ's city-transport export (data/raw/2026-10-20/JDF_mestske.zip, https://portal.cisjr.cz/pub/draha/mestske/JDF.zip, downloaded 21 Oct 2026); the shared 02_timetable.py reads it with the same version rule and records traction (T).
  • Fleet: config.fleet.carriers = Z-Group bus (seznam-autobusu 240), Transdev Slezsko (28), MDPO (338, incl. trolleybuses) and Transdev Morava (7546, for Hlučínsko's 24- buses) → 01c_fleet.py → data/derived/fleet.json. MDPO sends MDPO instead of a plate to the live map: live_plate maps it.
  • Looked-up plates (plates.json, 01b_plates.py): authoring/fix_plates.py rewrote their paints from all current layers (the shared parser keeps only the last layer: "logo dopravy MSK" instead of "bílá & červené logo Transdev & logo dopravy MSK").
  • Line 204: MDPO notice (https://www.mdpo.cz/pro-cestujici/aktuality-doprava/omezeni-provozu-mhd-na-ul-vrchni-v-opave): Vrchní street repairs, one-way 12–20 Oct, full closure 21–23 Oct; on 20 Oct 204 ran with replacement buses (trips 40xx, Citelis 150–157). Short works → 204 is modelled as the regular trolleybus line. Its trolleybuses are designed by authoring/fill_designed.py (config fill_designed): MDPO trolleybuses not in service 15:30–17:30. The shared roster cannot design them (it only takes vehicles that ran the area's lines that day, or fleet numbers with an NN- prefix).
  • Overhead wire (22 Oct 2026): authoring/offwire.py checks each trolleybus route against the OSM wire, the same test as ../praha-trolleybus/scripts/03b_offwire.py (10 m samples, 25 m tolerance, a section is off-wire under a third), with the OSM route relations as the path because the JDF has no shapes. Raw data in data/raw/osm/2026-10-22/ (Overpass, Accept: application/json; overpass-api.de often answers 502, maps.mail.ru's mirror worked): trolley_wire.json = way["trolley_wire"="yes"](49.88,17.78,50.00,18.02); out geom tags;, trolleybus_routes.json = relation["route"="trolleybus"](49.90,17.83,49.98,17.98); out geom;, stops.json = bus_stop / platform / stop_position nodes in the same box, out body;. Result (data/derived/offwire.json, a table on each trolleybus card): 201–203 under the wire all the way; 204 off it Ratibořská – Vrchní – U Cukrovaru – Pekařská – Divadlo (2.0 km without wire) and Kylešovice, Bílovecká – Hlavní – škola (0.6 km in a straight line between the stops, about 1 km by the JDF's whole-km distances). Both 204 relations are complete=no and stop at Bílovecká: the script adds a missing end from the line's timetable stops when every OSM platform of those stops is over 60 m from the wire (here: none within 100 m). Cross-check: Wikipedia (Trolejbusová doprava v Opavě) marks škola, Hlavní, Vrchní, U Cukrovaru and Pekařská as partial-trolleybus stops, and 202's Malé Hoštice trip (once a day, not modelled).
  • Auxiliary drive: the register's vehicle pages (data/raw/2026-10-22/vuz/_.html, seznam-autobusu.cz/vuz/, fetched with the shared browser headers) give "alternativní pohon": baterie on 32 Tr 314–323, dieselagregát on 26 Tr 305–306, nothing on the rest. authoring/alt_drive.py writes alt onto fleet.json and the roster (the page shows "+ battery" / "+ diesel generator" after the type), and fill_designed takes only those for 204 (alt, battery first). On 20 Oct 317, 318, 320, 321, 322 and 316 ran 202/203 in the window, so 204 has 315, 314, 319, 323 (battery) and 305 (diesel generator).
  • Off-wire objects (22 Oct, fleet desk, pak main 684fc09): VZ-DPOpava-32Tr-mdpobilocervena-32Tr_battery (battery) for 314–323 and VZ-DPOpava-26Tr-mdpobilocervena-26TrDG (diesel) for 305–306, chosen by the pak rules '32 Tr battery' / '26 Tr diesel generator' (the shared 06 matches a vehicle's alt with its type). 324–328 stay electric (the register gives them no alternative drive; the desk notes the register's attribute isn't always filled in).
  • The save's old ODIS lines (800…, incl. an electric-only 204) predate the fork and are being rebuilt by the owner (22 Oct): nothing on this sheet is fitted to them.
  • Pak: fleet desk answer of 21 Oct (via the parent session): Z-Group white + kraj logo = bila; Hlučínsko = red ring bilacervenykruh or transdev; MDPO = VZ-DPOpava-bus / VZ-DPOpava-trolleybus (advert wraps in plain livery, final); SOR BNG 10,5 loan = permanent stand-in VZ-MHDHavirov-SOR_BNG_10_5-havirovmodra. The desk promised VZ-ODIS-SOR_CNG_12-bilacervenykruh (not needed on this roster after the paint fix).

Decisions (21 Oct 2026)

  • Network state: short diversions → regular route (regular_not_diversion): 241, 243, 262 (Psychiatrická léčebna, to 23 Nov), 253 (Služovice, obecní úřad, to 31 Oct), 256, 275, 284 (Dolní Benešov, Zábřeh, Mexiko, 16–21 Oct). Diversions to the end of the timetable year run as they are: 234, 242, 261 (Březová, střed) and 263, 264.
  • Variants (each covers stops no other line serves): 236 Podlesí, 237 Dvorce, 241 Lipina, 245 zámek, 249 Tábor, 250 Krnov / Holasovice, 253 Vrbka, 260 Heřmanice, 262 Skřipov, 293 ŽD, 920 Rýmařov / Opava (285 Dobroslavice was one until 25 Oct: all its stops are on 285, so it is listed on the 285 card now). One-trip deviations are taken with stops_via (216 OPAVIA, 217 Otice, 232 Hvězdová, 238 Loučky, 241 Zálužné, 252 Albertovec, 261 Hladké Životice / Gručovice, 262 Slatina, 267 Jezdkovice, 678 Bravantice Lidl). terminus_alias joins Opava, Jaktař škola (town JDF) and Opava, Jaktař, škola (regional).
  • Hand-set after review: 204 (three slots :00/:24/:36), 217 (hourly :04/:25, rt 20), 247 (15:00/17:00), 248 (hourly :53/:45), 261 (every 120, 14:22/14:35), 920 (every 120 Šumperk / Rýmařov alternating at :17, Opava variant 13:24/15:02).
  • Owner checks on the page: 204 as trolleybus (with auxiliary-drive trolleybuses) vs as it ran; 217 Divadlo trips; regular routes of 241/243/262; the 920 split. Added 25 Oct (below): 241/252/262 with fewer buses, 215/233/264 with one bus, 248 every 120. The shared 06 writes the number of checks as a word up to seven, so this sheet is at its limit.

Hlučínsko lines at sad B. Němcové, checked by hand (25 Oct 2026)

The owner asked for 281, 282, 290 and 292 to be checked step by step ("that's the point of you doing it, so you'll not produce incorrect results"). Their 21 Oct rows were automatic fits and all failed Step 4. The check is in method-check-2026-10-25.md.

  • Outstation lines: the buses sleep at the far ends (Hať garage, Třebom, Strahovice, Chuchelná, Píšť), so the afternoon sends more buses out of Ostrava than come back (19 out, 12 in at sad B. Němcové, 14:00–17:00). Each line is timed both ways at its inbound rate; the extra peak departures (buses going home) are off the pattern. v7 had them on B lines with empty returns (5 more buses); the owner: "Unless those empty runs add something really needed and specific, don't add them just for frequency". They served no stop of their own, so v8 dropped them.
  • 290: every 120 with four slots each way, exact (Ostrava 14:44, 14:59, 15:24, 15:49; Hať 14:47, 15:31, 16:07, 16:37).
  • 281: Třebom :58, timed so the bus passes Kobeřice at :13 (was Třebom :16, 15–20 min off); Ostrava :18 unchanged.
  • 282: every 120 (Píšť :55 is exactly every 2 h; Ostrava :53), model hour set to 16:00 (anchor) because its busiest stretch was the Hlučín – Závada shuttles.
  • 292: every 80, Ostrava 14:33 / 15:53, back timed at Hlučín :59 (the trips start at Píšť, Chuchelná or Strahovice).
  • Needs 9 → 10. config.keep_roster (data/published/roster_v6.json, new opt-in step in the shared 04_roster) keeps the published buses where they are still seen on their line: 281 24-0004 → 24-0002, 290 + 24-1111, nothing else moves. 290: the owner chose the four exact slots and 4 buses over every 40 with 3 (25 Oct 2026: "290 uses more buses").

The rest of the sheet, checked by hand (25 Oct 2026)

Every leg the regional Step 4 audit (../audit/audit-regional-2026-10-25.md) listed as failing was checked by hand with the same method; the reading and the candidates are in method-check-2026-10-25.md ("The rest of the sheet"). All rows are hand-set now (no at: 'auto' on the changed lines).

  • Set exact in the model window (every leg passes Step 4): 212, 216, 218, 219 every 120 with three or four slots; 220 hourly with two; 238 every 90; 240, 242 every 120 with two; 245 every 80; 249 every 120; 270 every 80; 293 every 96 (every 120 would leave Hlučín on 293 ŽD's 13:05); 678 every 144 (every 120 needs 3 buses for the same two departures).
  • Model hour set by hand to 16:00 (anchor), as 282: 075 (its :02 / :22 extras made 13:40–15:40 the busiest) and 555 (two Bohumín, Nový Bohumín short trips did). 285 is timed in its own afternoon (window 13:00–15:00): its Štítina trips end at 15:50.
  • Inbound rate (outstation, owner rule of 25 Oct): 236 (buses sleep at Budišov), 263 (at Pustá Polom), 284 (its buses go on from Štěpánkovice); the extra departures are off the pattern.
  • Variants listed, not folded in: 242's express pair via Jelenice (Vítkov 14:30, Opava 14:45; via Branka u Opavy); 285 Dobroslavice is no longer a line (all its stops are on 285, and it counted 285's own trips past Dobroslavice).
  • Kept approximate: 251 and 248 hourly (a narrow window must not halve an hourly line), 254 every 90 (no pattern passes both ways), 217 and 920 (their owner checks).
  • Buses: 215, 233, 264 +1 (exact departures one bus cannot chain), 241 +2, 252 +1, 262 +1, 285 +1 and 285 Dobroslavice −1: 126 → 133. The cheaper options are on the owner checks (241/252/262 −4, 215/233/264 −3, 248 −1).

Authoring and review (authoring/)

  • opa_spec.py → config.lines (+ opa_notes.py); opa_config.py → everything else (pak rules, texts, areas, network state, owner checks, reasons, area_home, fill_designed, artifact_url). Both idempotent.
  • ends.py, dump.py, review.py, view.py, suggest.py, divcmp.py: review tools pointed at this folder.

How to rebuild

From ../odis-regional-hko/scripts:

  1. python -X utf8 run.py --base odis-regional-opavsko
  2. (automatic: run.py runs config.post_roster after 04_roster: authoring/offwire.py (wire check), authoring/alt_drive.py (battery / diesel marks), authoring/fill_designed.py (204's trolleybuses). For a new wire state, fetch the three Overpass files into a new data/raw/osm// and pass --osm or change the default.)
  3. python -X utf8 06_render.py --base odis-regional-opavsko
  4. Republish to artifact_url.

After a new 01b plate lookup, run authoring/fix_plates.py once before step 1's 04/06.