oddb2xml 3.0.30 → 3.0.32

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: dd5aba5cd43be1930a5499b102b5a1921454d422de9aa13de227e92ad1b5cae6
-  data.tar.gz: 1861fc5dba9e46179aaa4f69e64fa6e10288aeafff677ec591156926a54e8633
+  metadata.gz: 6526c05c2efd31be9fff9279dd24e817a3ae94363c9b30d8286dbef0510596d5
+  data.tar.gz: c8c835826c01e71719efbef0c73cb0f4805037dac8a7d89dcd82a7d58df41deb
 SHA512:
-  metadata.gz: eb67bef331b97a7f3b347d342c9c2c8128fe9d53154363b89d3868d97d61d2d6469568aa78fb597129f33687af68c2ef38ecb227a6e67d807b8e06df220e4694
-  data.tar.gz: b4835b7045e30b1e24a84afd96c499db2b09fe0ec0175ea8b8e82adfe9a7b4a096237e6be6879cb9636f9608a0161346dda37b1606abf38d62762d858375b056
+  metadata.gz: 8358ad2fd07b3cad7d4e87c7b364fa8ea0a3f68742c547d9781dbf61614c6a2d85b3f95bd5ee8f3c7d0aa48bb08b8e2b45eb279382778e25a5333f6928440fed
+  data.tar.gz: e332eb5a017afd15dc78663d12f2956469ec2ed4e9de4742b28771f84dd10983593787c2a611327c130202ed82f34c56c730b75c24e9eb4ad49912cdb3cb5519

data/CLAUDE.md

@@ -6,31 +6,6 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
 
 oddb2xml is a Ruby gem that downloads Swiss pharmaceutical data from 10+ sources (Swissmedic, BAG, Refdata, ZurRose, EPha, etc.), parses multiple formats (XML, XLSX, CSV, SOAP, fixed-width DAT), merges/deduplicates them, and generates standardized XML/DAT output files for healthcare systems. It also supports the Elexis EHR Artikelstamm format.
 
-## Common Commands
-
-```bash
-# Install dependencies
-bundle install
-
-# Run full test suite
-bundle exec rake spec
-
-# Run a single test file
-bundle exec rspec spec/builder_spec.rb
-
-# Run a single test by line number
-bundle exec rspec spec/builder_spec.rb:42
-
-# Lint with StandardRB
-bundle exec standardrb
-
-# Auto-fix lint issues
-bundle exec standardrb --fix
-
-# Build the gem
-bundle exec rake build
-```
-
 ## Architecture
 
 The system follows a **download → extract → build → compress** pipeline:
@@ -47,7 +22,7 @@ The system follows a **download → extract → build → compress** pipeline:
 
 6. **Compressor** (`lib/oddb2xml/compressor.rb`) — Optional ZIP/TAR.GZ output compression.
 
-7. **FHIR support** (`lib/oddb2xml/fhir_support.rb`) — Self-contained module providing `FhirDownloader` and FHIR NDJSON parsing. Activated via `--fhir` (or `--fhir-url=<URL>`). Downloads per-language NDJSON files (`foph-sl-export-latest-{de,fr,it}.ndjson`) from `epl.bag.admin.ch` to populate French and Italian product names/descriptions. Maps legal status codes `756005022007` and `756005022008` to Swissmedic category D. Reads the BAG **Indikationscode** (`XXXXX.NN`) from the explicit `indicationCode` extension on each `RegulatedAuthorization.indication[].extension[regulatedAuthorization-limitation]` (BAG SL FHIR export >= v2.0.5; handled from 3.0.10). The BAG changelog states the limitation code (`ClinicalUseDefinition.id`) and the indication code are **independent** fields, so the older derivation — combining each indication CUD's `.NN` id-suffix with the reimbursement RA's `FOPHDossierNumber` — is kept only as a fallback for feeds lacking the extension. Exposed as `item[:indication_codes]` and per-package `:indication_codes` (each entry a `{code:, cud_id:, text:}` hash, where `cud_id` is the `limitationIndication` CUD reference used to resolve the text). From 3.0.7 onwards, `Builder#build_product` emits one `<INDICATION_CODE code="XXXXX.NN" cud_id="DRUG.NN">limitation text</INDICATION_CODE>` child per indication on every `<PRD>` in `oddb_product.xml`; live feed numbers: 539 products / 1,293 codes / 100 % with non-empty indication text. Mandatory on prescriptions/invoices for SL price-model drugs from 2026-07-01 — see issue [#113](https://github.com/zdavatz/oddb2xml/issues/113). **Limitation texts** (3.0.8 onwards): the `regulatedAuthorization-limitation` extension has no inline `limitationText` in the live BAG feed — it carries a `limitationIndication` reference to a `ClinicalUseDefinition` whose `indication.diseaseSymptomProcedure.concept.text` is the actual text. The parser stores the ref as `cud_ref` on each Limitation, `Bundle#cud_text_by_id` resolves DE, and `merge_language` propagates FR/IT from the per-language NDJSON files via the same CUD id. Coverage on the live feed jumped from 0 / 9'108 to 9'108 / 9'108 (issue [#116](https://github.com/zdavatz/oddb2xml/issues/116)). **Limitation code / LIMNAMEBAG** (3.0.12 onwards): FHIR has no native BAG limitation code (LIMCD), so `create_limitations_for_package` sets `LimitationCode = cud_ref` (the `limitationIndication` CUD id) instead of `""`. Without this, every FHIR limitation shared an empty `:code`; `Builder#build_artikelstamm` groups its `<LIMITATIONS>` section by code, so all of them collapsed into a single `<LIMITATION>` with an empty `<LIMNAMEBAG>` and only one text survived. Using the CUD id as the key makes each distinct limitation emit and be referenced from its `<PRODUCT>`. The downstream `bin/check_artikelstamm` (`semantic_check.rb`) also crashed on the lone-element output because Ox `:hash_no_attrs` collapses a one-child section into a Hash (and an empty one into nil) — `SemanticCheckXML#get_items` now normalises every section to an Array. **v6 Artikelstamm / per-article INDC (3.0.26 onwards):** `--artikelstamm` now emits the **Elexis Artikelstamm v6** format (namespace `http://elexis.ch/Elexis_Artikelstamm_v6`, file `artikelstamm_DDMMYYYY_v6.xml`/`.csv`, validated against the bundled `Elexis_Artikelstamm_v6.xsd`) — replacing v5. The new piece is a per-`<ITEM>` `<ARTSL>` block carrying the BAG Indikationscodes (issue [#113](https://github.com/zdavatz/oddb2xml/issues/113)): `<PM>true</PM>` plus one `<ARTLIM>` per limitation with `<LIMCD>` (= `cud_ref`, the BAG limitation code), `<INDCD>` (the `XXXXX.NN` indication code from the `indicationCode` extension), and `<VDAT>`/`<VTDAT>` (period start/end). To feed it, `create_limitations_for_package` now also carries `IndicationCode` (→ per-package `:indcd`) and `ValidThruDate` (→ `:vtdate`) on each limitation; `Builder#append_artsl`/`elexis_datetime` emit one `<ARTLIM>` per limitation that has a non-empty `:indcd` (so non-price-model items get no `<ARTSL>`). `PM` is always `true` here because the indication code is required only for SL price-model drugs, which is exactly the set of items that reach this block. The bundled `Elexis_Artikelstamm_v6.xsd` is the canonical MEDEVIT schema extended with oddb2xml's historical Italian elements (`DSCRI` on PRODUCT/LIMITATION/ITEM, `DOSAGE_FORMI` on ITEM) so the output still validates. The legacy `--no-fhir` path emits no `<ARTSL>` (no FHIR limitations). **Per-article INDC in the `-e`/`-b` feeds (3.0.27 onwards):** the same indication codes are also emitted per `<ART>` in `oddb_article.xml` via `Builder#append_indication_codes` — one `<INDICATION_CODE code="XXXXX.NN" cud_id="DRUG.NN" limcd="DRUG.NN" vdat=… vtdat=…>limitation text</INDICATION_CODE>` per limitation that carries an `:indcd` (sourced from the article's `pac[:limitations]`). This reuses the flat `<INDICATION_CODE>` element already emitted on `<PRD>` in `oddb_product.xml` (since 3.0.7), enriched with the limitation code and validity dates so it carries the full `<ARTSL>` payload. The bundled `oddb2xml.xsd` now defines `<INDICATION_CODE>` (simpleContent + `code`/`cud_id`/`limcd`/`vdat`/`vtdat` attributes) and references it from both `<ART>` and `<PRD>` — the `<PRD>` reference had been missing since the 3.0.7 addition, so the FHIR product feed only validated once this was added. **Legacy v5 opt-in (3.0.28 onwards):** `--artikelstamm-v5` additionally emits the older Artikelstamm v5 file alongside the v6 one (for consumers not yet migrated). It implies `--artikelstamm`; `Options.parse` sets `@opts[:artikelstamm] = true`. `build_artikelstamm` takes a `version:` keyword (default `6`) that drives the namespace (`Elexis_Artikelstamm_v#{version}`) and CSV/XML filename suffix; `build_artikelstamm_v5` calls it with `version: 5`. The only content difference is that the v5 path skips `append_artsl` (the v5 schema has no `<ARTSL>` element). The CLI `files` map adds an `:artikelstamm_v5` entry (→ `build_artikelstamm_v5` via the `to_xml` subject dispatch) and validates it against the bundled `Elexis_Artikelstamm_v5.xsd`. Both files come from one build — the `@prepared` guard means the shared download/prepare phase runs once.
+7. **FHIR support** (`lib/oddb2xml/fhir_support.rb`) — Self-contained module providing `FhirDownloader` and FHIR NDJSON parsing. Activated via `--fhir` (or `--fhir-url=<URL>`). Downloads per-language NDJSON files (`foph-sl-publication-latest-{de,fr,it}.ndjson`) from `epl.bag.admin.ch/static/sl/publication/fhir` (BAG moved the export there on 24.08.2026; the old `/static/fhir/foph-sl-export-*` answers 404 for `-latest-` while the old dated snapshots remain, so the failure looks like an import that does nothing) to populate French and Italian product names/descriptions. Maps legal status codes `756005022007` and `756005022008` to Swissmedic category D. Reads the BAG **Indikationscode** (`XXXXX.NN`) from the explicit `indicationCode` extension on each `RegulatedAuthorization.indication[].extension[regulatedAuthorization-limitation]` (BAG SL FHIR export >= v2.0.5; handled from 3.0.10). The BAG changelog states the limitation code (`ClinicalUseDefinition.id`) and the indication code are **independent** fields, so the older derivation — combining each indication CUD's `.NN` id-suffix with the reimbursement RA's `FOPHDossierNumber` — is kept only as a fallback for feeds lacking the extension. Exposed as `item[:indication_codes]` and per-package `:indication_codes` (each entry a `{code:, cud_id:, text:}` hash, where `cud_id` is the `limitationIndication` CUD reference used to resolve the text). From 3.0.7 onwards, `Builder#build_product` emits one `<INDICATION_CODE code="XXXXX.NN" cud_id="DRUG.NN">limitation text</INDICATION_CODE>` child per indication on every `<PRD>` in `oddb_product.xml`; live feed numbers: 539 products / 1,293 codes / 100 % with non-empty indication text. Mandatory on prescriptions/invoices for SL price-model drugs from 2026-07-01 — see issue [#113](https://github.com/zdavatz/oddb2xml/issues/113). **Limitation texts** (3.0.8 onwards): the `regulatedAuthorization-limitation` extension has no inline `limitationText` in the live BAG feed — it carries a `limitationIndication` reference to a `ClinicalUseDefinition` whose `indication.diseaseSymptomProcedure.concept.text` is the actual text. The parser stores the ref as `cud_ref` on each Limitation, `Bundle#cud_text_by_id` resolves DE, and `merge_language` propagates FR/IT from the per-language NDJSON files via the same CUD id. Coverage on the live feed jumped from 0 / 9'108 to 9'108 / 9'108 (issue [#116](https://github.com/zdavatz/oddb2xml/issues/116)). **Limitation code / LIMNAMEBAG** (3.0.12 onwards): FHIR has no native BAG limitation code (LIMCD), so `create_limitations_for_package` sets `LimitationCode = cud_ref` (the `limitationIndication` CUD id) instead of `""`. Without this, every FHIR limitation shared an empty `:code`; `Builder#build_artikelstamm` groups its `<LIMITATIONS>` section by code, so all of them collapsed into a single `<LIMITATION>` with an empty `<LIMNAMEBAG>` and only one text survived. Using the CUD id as the key makes each distinct limitation emit and be referenced from its `<PRODUCT>`. The downstream `bin/check_artikelstamm` (`semantic_check.rb`) also crashed on the lone-element output because Ox `:hash_no_attrs` collapses a one-child section into a Hash (and an empty one into nil) — `SemanticCheckXML#get_items` now normalises every section to an Array. **v6 Artikelstamm / per-article INDC (3.0.26 onwards):** `--artikelstamm` now emits the **Elexis Artikelstamm v6** format (namespace `http://elexis.ch/Elexis_Artikelstamm_v6`, file `artikelstamm_DDMMYYYY_v6.xml`/`.csv`, validated against the bundled `Elexis_Artikelstamm_v6.xsd`) — replacing v5. The new piece is a per-`<ITEM>` `<ARTSL>` block carrying the BAG Indikationscodes (issue [#113](https://github.com/zdavatz/oddb2xml/issues/113)): `<PM>true</PM>` plus one `<ARTLIM>` per limitation with `<LIMCD>` (= `cud_ref`, the BAG limitation code), `<INDCD>` (the `XXXXX.NN` indication code from the `indicationCode` extension), and `<VDAT>`/`<VTDAT>` (period start/end). To feed it, `create_limitations_for_package` now also carries `IndicationCode` (→ per-package `:indcd`) and `ValidThruDate` (→ `:vtdate`) on each limitation; `Builder#append_artsl`/`elexis_datetime` emit one `<ARTLIM>` per limitation that has a non-empty `:indcd` (so non-price-model items get no `<ARTSL>`). `PM` is always `true` here because the indication code is required only for SL price-model drugs, which is exactly the set of items that reach this block. The bundled `Elexis_Artikelstamm_v6.xsd` is the canonical MEDEVIT schema extended with oddb2xml's historical Italian elements (`DSCRI` on PRODUCT/LIMITATION/ITEM, `DOSAGE_FORMI` on ITEM) so the output still validates. The legacy `--no-fhir` path emits no `<ARTSL>` (no FHIR limitations). **Per-article INDC in the `-e`/`-b` feeds (3.0.27 onwards):** the same indication codes are also emitted per `<ART>` in `oddb_article.xml` via `Builder#append_indication_codes` — one `<INDICATION_CODE code="XXXXX.NN" cud_id="DRUG.NN" limcd="DRUG.NN" vdat=… vtdat=…>limitation text</INDICATION_CODE>` per limitation that carries an `:indcd` (sourced from the article's `pac[:limitations]`). This reuses the flat `<INDICATION_CODE>` element already emitted on `<PRD>` in `oddb_product.xml` (since 3.0.7), enriched with the limitation code and validity dates so it carries the full `<ARTSL>` payload. The bundled `oddb2xml.xsd` now defines `<INDICATION_CODE>` (simpleContent + `code`/`cud_id`/`limcd`/`vdat`/`vtdat` attributes) and references it from both `<ART>` and `<PRD>` — the `<PRD>` reference had been missing since the 3.0.7 addition, so the FHIR product feed only validated once this was added. **Legacy v5 opt-in (3.0.28 onwards):** `--artikelstamm-v5` additionally emits the older Artikelstamm v5 file alongside the v6 one (for consumers not yet migrated). It implies `--artikelstamm`; `Options.parse` sets `@opts[:artikelstamm] = true`. `build_artikelstamm` takes a `version:` keyword (default `6`) that drives the namespace (`Elexis_Artikelstamm_v#{version}`) and CSV/XML filename suffix; `build_artikelstamm_v5` calls it with `version: 5`. The only content difference is that the v5 path skips `append_artsl` (the v5 schema has no `<ARTSL>` element). The CLI `files` map adds an `:artikelstamm_v5` entry (→ `build_artikelstamm_v5` via the `to_xml` subject dispatch) and validates it against the bundled `Elexis_Artikelstamm_v5.xsd`. Both files come from one build — the `@prepared` guard means the shared download/prepare phase runs once. **Selbstbehalt / costShare (3.0.31 onwards):** the `reimbursementSL` extension's `costShare` valueInteger is the BAG Selbstbehalt percentage; it maps to the legacy `FlagSB` semantics of the old BAG `Preparations.xml` — `40` → `item[:deductible] = "Y"` (raised Selbstbehalt: `<SLOPLUS>1</SLOPLUS>` in `oddb_article.xml`, `<DEDUCTIBLE>40</DEDUCTIBLE>` in the Artikelstamm), `10` → `"N"` (ordinary Selbstbehalt: `SLOPLUS` 2 / `DEDUCTIBLE` 10), `20` → `deductible20 = "Y"` (transitional rate, issue #81). Until 3.0.31 the mapping was inverted (`10` → `"Y"`), so **every** SL article claimed a 40 % Selbstbehalt. On the live feed only 385 of 10'388 packages (222 preparations) carry `costShare` 40; the value never mixes within one preparation, so keeping the flag at sequence level is safe.
 
 8. **Refdata cleanup** (`lib/oddb2xml/refdata_cleanup.rb`) — Compensates for known data-quality issues in upstream Refdata.Articles.xml before they reach the output. Each fix is guarded by a Swissmedic-side heuristic (e.g. comma in `substance_swissmedic` to distinguish mono products from real combinations). Currently fixes (a) the doubled-dose template bug (`X mg / X mg / Stk`, `fix_double_dose`, guarded by `single_substance?`); (b) the spelled-out German galenic form `Retardtabletten` → house-style abbreviation `Ret Tabl` (`normalize_galenic_form` / `GALENIC_NORMALISATIONS`, issue #112 case #13, e.g. RINVOQ — a narrow word-boundary substitution that leaves legitimate brand suffixes like `TRAMAL retard` and Mepha's `Lactab` untouched); and (c) dose info Refdata dropped from `<FullName>`, sourced from the Swissmedic composition string `pack[:composition_swissmedic]` — `fix_missing_combo_dose` (#6, appends a combination's 2nd component strength), `fix_missing_dose` (#4, inserts a mono product's missing strength before the pack count), `fix_missing_volume` (#7, appends an injectable's per-pen volume); and (d) 50-char-truncation repairs — `fix_truncated_metoject` (#1, rebuilds METOJECT Autoinjektor names from the intact `<brand> Autoinjektor <dose>/<vol>` prefix + Swissmedic `size`, localised DE/FR/IT) and `fix_truncated_volume_unit` (#3, restores the cut `ml` of the VERACTIV Vitamin D3 drops). The (c) and (d) fixes are scoped to explicit IKSNR allow-lists (`COMBO_DOSE_IKSNR`/`MISSING_DOSE_IKSNR`/`MISSING_VOLUME_IKSNR`/`METOJECT_IKSNR`/`VERACTIV_VITD3_IKSNR`): a dry run proved a blanket heuristic mis-fires on hundreds of legitimate names (sodium counter-ion doses, strength-less phyto/powder products, concentration names like `CIMZIA 200 mg/ml`), so only catalogued registrations are touched — add an IKSNR to grow coverage. Called from `Builder#apply_refdata_description_cleanups!` at the start of `prepare_articles`. See GitHub issue #112 for the catalogue.
 
@@ -70,12 +45,16 @@ YAML files in `data/` provide manual overrides and mappings: `article_overrides.
 
 These scripts run the public download server at `https://mediupdatexml.oddb.org` (Apache on this host) and are **not** part of the gem itself.
 
+- **`get_transfer.sh`** (cron: `30 0 * * * zdavatz`, restored 2026-08-02) — mirrors the ZurRose IGM11 **Vollstamm** article master to `$DATA_DIR/TRANSFER.ZIP` (`/home/zdavatz/software/get_transfer`, the `$GET_TRANSFER_ZIP` that `run_oddb2xml.sh` seeds into `downloads/`). Downloads `transfer.dat`, validates it, and zips it with the entry named **lowercase `transfer.dat`** — `ZurroseDownloader` (`downloader.rb:226`) unzips into `downloads/` and opens that exact name. **Two things had gone wrong at once.** (a) The script lived *only* in `/home/zdavatz/software/get_transfer` on the host HIN deleted, so it vanished with the box and the cron line — guarded by `[ -x … ]` — went **silently** dead; it is now tracked in `scripts/` with that path a symlink, so `setup_new_server.sh` restores it. (b) Its source URL died independently around mid-July 2026: the old `zurrose.com/fileadmin/main/lib/download.php?file=…` now 301s to the marketing homepage. Current URL is `https://www.zurrose.ch/sites/default/files/media/downloads/medi/produktupdate/igm11/de/transfer.dat` (override with `TRANSFER_URL`). Consequence of both: `pillbox.oddb.org/TRANSFER.ZIP` — which is only a mirror *of this job* — froze at its 2026-07-16 copy. **Validation** is a record-shape check, not a size check: the file is fixed-width IGM11, every record exactly **98** bytes (179'856 records on the 2026-08-02 file), and fewer than `MIN_LINES` (100'000) records or any off-length record is rejected. A ZurRose error page passes a size check and fails this one. **Last-good semantics:** the download goes to a scratch dir and is published by `mv` only after validation, so a 404 or an HTML page leaves the previous `TRANSFER.ZIP` untouched — the 01:00 build reads that zip unconditionally, and half a file there aborts the whole nightly run at the unzip step. ZurRose refreshes the master roughly every 14 days, so most runs legitimately fetch an unchanged file (logged, not an error).
 - **`run_oddb2xml.sh`** — nightly build driver (cron: `0 1 * * * zdavatz`). Downloads the upstream sources **once**, then builds the `-b`/firstbase feed at price increments `45/50/55` plus `default` (no increment) into `$OUT_DIR` (`/home/zdavatz/oddb2xml`, one subdir each). The shared `downloads/` cache and transient zip live in `$BUILD_DIR` (`<OUT_DIR>-build`), **outside** `$OUT_DIR` so the transfer never uploads the multi-hundred-MB cache. Final step ("2b") regenerates the landing page. Each `oddb2xml` invocation is wrapped in `run_with_retry` (default **3 attempts, 120 s apart**, tunable via `ODDB2XML_RETRIES`/`ODDB2XML_RETRY_DELAY`): a transient upstream download failure (e.g. Swissmedic resetting the connection, `Errno::ECONNRESET`) previously aborted the whole `set -e` run 14 s in and left the feeds a day stale, so it now retries before giving up; a genuine repeated failure still stops the run. **Firstbase (GS1 NONPHARMA) last-good fallback (3.0.29 onwards):** the GS1 `GetFirstbaseHealthcare` export (`id.gs1.ch/01/07612345000961` → `apitools.gs1.ch`) has been answering `403 - Forbidden`, which blanked `firstbase.csv` and dropped **every** NONPHARMA article from the `-b` feed (landing page then showed `NONPHARMA = 0 − 1 = −1`). The script keeps the last successful `firstbase.csv` in a persistent cache `$FIRSTBASE_CACHE` (default `<OUT_DIR>-state/firstbase.csv`) **outside** `$BUILD_DIR` so it survives the nightly `rm -rf`, seeds it into `downloads/` before the build, and refreshes it after a successful download. The gem side (`FirstbaseDownloader#download`, rewritten in 3.0.29) makes the seed usable: it still attempts the live GS1 fetch on the first (downloading) build, but only overwrites `firstbase.csv` when the response is a real non-empty CSV (`firstbase_csv?` rejects HTML/`403 - Forbidden`/empty bodies and open-uri exceptions), otherwise it **keeps the existing seeded file** instead of the old `"w+"` truncate-to-zero. A recovered GS1 therefore refreshes the data automatically; while GS1 is down the feed serves yesterday's (last-good) NONPHARMA rather than nothing. `generate_index_html.sh` also guards the NONPHARMA count so an empty CSV renders `—` (not `−1`). **ZurRose/pillbox last-good seed + first-build `--skip-download` (2026-07-07):** the actual reason the nightly feed went stale after 2026-07-03 was **not** firstbase (a firstbase failure is non-fatal) but **ZurRose**: `pillbox.oddb.org:80` started refusing connections around 01:00 every night from 2026-07-04 on, so `ZurroseDownloader` wrote a 0-byte `transfer.zip`, `unzip` produced no `transfer.dat`, and `downloader.rb:221` raised `Errno::ENOENT`, aborting the whole `set -e` run (all three retries hit the same dead pillbox). Fix is **script-only, no gem release**: `get_transfer.sh` (cron 00:30) already mirrors the ZurRose zip locally at `$GET_TRANSFER_ZIP` (`/home/zdavatz/software/get_transfer/TRANSFER.ZIP`), so the **first** `build_one` now runs a `first_build_attempt` helper that (a) `seed_downloads` — resets `downloads/` to contain only that seeded `transfer.zip` — then (b) runs `oddb2xml --skip-download -b -c zip`. The `--skip-download` flag makes `Cli#run` **skip the `FileUtils.rm_f(Dir.glob(DOWNLOADS/*))` wipe (cli.rb:50)** that used to delete the seed, so `Oddb2xml.skip_download` reuses the seeded `transfer.zip` for ZurRose (no pillbox fetch) while `download_as` still fetches **every other source fresh** (Swissmedic, Refdata, BAG, FHIR NDJSON, Weleda/WALA, and firstbase — proven: `--skip-download` only reuses files already present in `downloads/`, it does not suppress downloading absent ones). Seeding happens **per attempt** inside `run_with_retry`, so a retry restarts from a clean cache (clears partials) while keeping the seed. **Consequence for firstbase:** it is now downloaded **once** by the first build and reused by the `45/50/55`/artikelstamm increments via `--skip-download` — and it is **no longer pre-seeded into `downloads/`** from `$FIRSTBASE_CACHE` (under `--skip-download` a *present* `firstbase.csv` would be reused verbatim and never refreshed, freezing NONPHARMA). GS1's `id.gs1.ch` route currently returns a real 151 MB CSV, so firstbase fetches fresh; `$FIRSTBASE_CACHE` is still refreshed from the fresh file after the build for reference/recovery. The 3.0.29 gem-side `firstbase_csv?`/keep-existing logic is retained but only engages on the non-`--skip-download` path.
 - **`generate_index_html.sh DOCROOT [FIRSTBASE_CSV]`** — single source of truth for the landing page. Writes `index.html` + a self-contained `logo.svg` **atomically** (temp + `mv`, so either owner — root from setup, `zdavatz` from cron — can refresh it). Computes live counts: PHARMA = `<SMNO>` count in `default/oddb_article.xml`, NONPHARMA = firstbase CSV rows − 1, total ART = `<ART ` count. Also runs **`visitor_stats.py`** and embeds its graph. Re-run standalone any time (it only reads already-built files); a separate cron line refreshes it **hourly** (`5 * * * * zdavatz`) so counts + graph stay current between nightly builds.
 - **`visitor_stats.py LOG_GLOB CACHE_DIR [DAYS]`** — emits the visitors/sessions/region graph as an inline-SVG HTML **fragment** (last `DAYS`, default 14): Besucher = distinct IPs/day, Sitzungen = 30-min-inactivity sessions per `(IP, User-Agent)`, plus a top-6 country breakdown by IP. Bots are filtered by User-Agent. Region lookup is **fully self-contained** — pure Python stdlib + the free **DB-IP country-lite CSV** (CC-BY, no licence key) cached in the build `downloads/` dir and refreshed monthly; **no apt package, no gem, no system GeoIP DB**. Prints nothing (page degrades to omitting the section) when the Apache log is unreadable or empty. Reading `/var/log/apache2` requires the cron user to be in the **`adm`** group (`sudo usermod -aG adm zdavatz`).
-- **`swissmedic_watch.sh`** — outage/block auto-recovery (cron: `*/30 * * * * zdavatz`). Since the Swissmedic platform migration (~2026-06-23, now a Swisscom-operated gateway), `www.swissmedic.ch` intermittently resets this host's automated connections **after the TLS handshake** (TCP RST), which aborts `run_oddb2xml.sh` under `set -e` and leaves the feeds stale (the block is host/IP- and client-fingerprint-sensitive: a real browser works, `curl`/`wget`/Ruby get reset, while other admin.ch hosts answer fine — so it is a WAF/bot rule, not an outage). The watcher polls Swissmedic with **oddb2xml's own client** (a Ruby `open-uri` canary on `listen_neu.html`); while blocked it is a silent no-op, and the moment it gets HTTP 200 it launches **one** build and emails. It fires **at most once per day** (stamp in `$STATE_DIR`, default `<OUT_DIR>-watch`, kept **outside** the wiped `$BUILD_DIR`), and skips when a build is already running or today's `default/oddb_article.xml` is already fresh. It exports `RBENV_VERSION=3.4.5` + the rbenv-shims PATH to match the nightly cron (the repo `.ruby-version` pins an uninstalled Ruby).
+- **`swissmedic_watch.sh`** — outage/block auto-recovery (cron: `*/30 * * * * zdavatz`). Since the Swissmedic platform migration (~2026-06-23, now a Swisscom-operated gateway), `www.swissmedic.ch` intermittently resets this host's automated connections **after the TLS handshake** (TCP RST), which aborts `run_oddb2xml.sh` under `set -e` and leaves the feeds stale (the block is host/IP- and client-fingerprint-sensitive: a real browser works, `curl`/`wget`/Ruby get reset, while other admin.ch hosts answer fine — so it is a WAF/bot rule, not an outage). The watcher polls Swissmedic with **oddb2xml's own client** (a Ruby `open-uri` canary on `listen_neu.html`); while blocked it is a silent no-op, and the moment it gets HTTP 200 it launches **one** build and emails. It fires **at most once per day** (stamp in `$STATE_DIR`, default `<OUT_DIR>-watch`, kept **outside** the wiped `$BUILD_DIR`), and skips when a build is already running or today's `default/oddb_article.xml` is already fresh. **PATH assembly (2026-07-28 onwards):** cron hands the watcher a minimal PATH, so it builds one explicitly — `/usr/local/bin` (where a root `gem install` under Debian's system Ruby puts the binstub) + the user gem bin dir (`$(ruby -e 'print Gem.user_dir')/bin`, where an *unprivileged* `gem install` silently lands instead, because `/var/lib/gems` is not writable) + rbenv shims **only when they exist** (the pre-2026-07 box; `RBENV_VERSION=3.4.5` is exported only in that case). The previous hardcoded rbenv-only PATH carried no `/usr/local/bin`, so the first watcher-triggered build on the rebuilt server died `oddb2xml: Kommando nicht gefunden` (exit 127) on all three retries before aborting. `run_oddb2xml.sh` step 1 extends PATH the same way right after its `gem install`, for the same reason.
 - **`transfer.sh`** — optional hand-off (scp) of `$OUT_DIR` to the HIN host; `SCP_DEST` is required-but-unset until the HIN host is known.
-- **`setup_mediupdatexml_web.sh`** — one-time root setup of the Apache vhost + initial page.
+- **`setup_new_server.sh`** (2026-07-28 onwards) — one-command rebuild of a bare Debian host into this download site, written after **HIN deleted the previous server**: everything that had accumulated by hand on that box, so a rebuild is not a day of archaeology. Root, idempotent. Installs the packages (`apache2` + `certbot`; **`ruby-full`** — Debian 13 ships Ruby 3.3, matching the repo `.ruby-version`, so the new host has **no rbenv**; `cron`, absent on a minimal image; `unzip`/`zip` for the ZurRose transfer and the build zip; `poppler-utils` for `pdftotext` in `tools/generate_bag_sl_group_prices.rb`; nokogiri/ffi build headers), adds `$RUN_USER` to **`adm`** (so `visitor_stats.py` may read `/var/log/apache2`), sets the home dir to **711** (so `www-data` can traverse to `$OUT_DIR` without listing it), `gem install oddb2xml`, creates the five directories (`$OUT_DIR`, `-build`, `-state`, `-watch` — the last two deliberately outside the nightly-wiped `-build` — and `$GET_TRANSFER_DIR`, symlinking the tracked `get_transfer.sh` into it), writes **`/etc/cron.d/mediupdatexml`** (the whole schedule in one re-creatable file rather than the old box's per-user crontab: `get_transfer` 00:30, nightly build 01:00, rust2xml Artikelstamm 03:00, aips2sqlite Fachinfos 04:30, landing page hourly at `:05`, Swissmedic watcher every 30 min) plus a logrotate rule for the cron logs, then hands over to `setup_aips2sqlite.sh` and `setup_mediupdatexml_web.sh`. Env: `RUN_USER`, `OUT_DIR`, `SKIP_WEB=1`, `SKIP_GEM=1`, `SKIP_AIPS=1`. **Deliberately out of scope** (gone with the old server, restored separately): an MTA for the cron `MAILTO`.
+- **`setup_mediupdatexml_web.sh`** — one-time root setup of the Apache vhost + initial page. Skips the Let's Encrypt step with a hint when the domain does not yet resolve in **public** DNS (checked via Cloudflare DoH, not the local resolver, which may hold a stale or negative cache).
+- **`setup_aips2sqlite.sh`** (2026-08-01 onwards) — provisions the **aips2sqlite** half of the site, restored after the server rebuild. The landing page links a whole `/aips2sqlite/` section (Fachinformationen as XML, the AmiKo `.db` files, `oddb2xml_swissmedic_sequences.csv`, `atc_codes_used_set.txt`); `setup_mediupdatexml_web.sh` writes the Apache alias `/aips2sqlite` → `<checkout>/jars/output`, but on the rebuilt host neither the JRE nor that directory existed, so **every link answered 403** — Apache falls back to the default `<Directory /> Require all denied` when the aliased path does not exist, so a *missing* target reads as *forbidden*, not *not found*. Root, idempotent: installs **`openjdk-21-jre-headless`** (the repo ships a prebuilt fat jar `jars/aips2sqlite.jar`, so a JRE suffices — no JDK/Gradle unless the jar is rebuilt; the only AWT user, `BarCode.java`/barcode4j, is off the Fachinfo path), clones the checkout if missing, creates `jars/output` + `jars/downloads` owned by `$RUN_USER`, and appends the 04:30 cron entry to `/etc/cron.d/mediupdatexml` if not already there (`SKIP_CRON=1` when called from `setup_new_server.sh`, which writes that line itself). Generating the data is a separate build-user step: `aips2sqlite/scripts/generate_aips_fi` (~1 h; `--lang=de --xml`, then `--lang=fr --xml --nodown`, then `--nodown --smsequence`, renaming `swiss_medic_sequences.csv` → `oddb2xml_swissmedic_sequences.csv`). `REFDATA_API_KEY` (developer.refdata.ch) is **optional** for this path: the Refdata Partner SOAP download it gates is consumed only by the Takeda partner export, and its failure is caught and logged, so the FI/sequences output is complete without it.
+- **`setup_rust2xml.sh`** — the build-user (never root) counterpart: installs the Rust toolchain via **rustup** — not Debian's `rustc`, because `rust2xml/scripts/run_artikelstamm.sh` sources `$HOME/.cargo/env` to find cargo under cron's minimal PATH — and runs `cargo build --release --bin rust2xml` in `$RUST2XML_DIR` (default `~/software/rust2xml`), feeding the 03:00 cron entry that publishes rust2xml's own Artikelstamm to `/artikelstamm/rust2xml/`. Idempotent: re-running updates the toolchain and rebuilds only if sources changed.
 
 Only the scripts are git-tracked; the generated `index.html`/`logo.svg` and the `downloads/` cache are not.
 
@@ -85,8 +64,3 @@ Only the scripts are git-tracked; the generated `index.html`/`logo.svg` and the
 - Test fixtures: `spec/data/` (sample files), `spec/fixtures/vcr_cassettes/` (recorded HTTP responses)
 - `spec/spec_helper.rb` defines test constants (GTINs) and configures VCR to avoid real HTTP calls during tests
 - CI runs on Ruby 3.0, 3.1, 3.2
-
-## Ruby Version
-
-- Minimum: Ruby >= 2.5.0 (gemspec)
-- Current development: Ruby 3.3.6 (`.ruby-version`)

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    oddb2xml (3.0.30)
+    oddb2xml (3.0.32)
       csv
       htmlentities
       httpi
@@ -89,10 +89,10 @@ GEM
     net-http-persistent (4.0.8)
       connection_pool (>= 2.2.4, < 4)
     nkf (0.2.0)
-    nokogiri (1.19.3)
+    nokogiri (1.19.4)
       mini_portile2 (~> 2.8.2)
       racc (~> 1.4)
-    nokogiri (1.19.3-arm64-darwin)
+    nokogiri (1.19.4-arm64-darwin)
       racc (~> 1.4)
     nori (2.7.1)
       bigdecimal
@@ -210,7 +210,7 @@ GEM
     webrobots (0.1.2)
     xml-simple (1.1.9)
       rexml
-    yard (0.9.43)
+    yard (0.9.44)
 
 PLATFORMS
   arm64-darwin-25

data/History.txt

@@ -1,3 +1,9 @@
+=== 3.0.32 / 26.08.2026
+* Bugfix (--fhir, the default): follow the BAG SL export to its new URL. BAG moved the publication on 24.08.2026 and announced it the next evening: /static/fhir/foph-sl-export-*.ndjson became /static/sl/publication/fhir/foph-sl-publication-*.ndjson. Only the -latest- alias is gone (the old dated snapshots stay in place), so the old URL failed by downloading nothing and reporting success rather than by crashing -- ch.oddb.org ran that way for twelve mornings with three HTTP 404 lines in a log as the only trace. Without the NDJSON there are no French/Italian names, no SL prices, no limitations and no Indikationscodes. Per the announcement there is no language-less default any more; every URL has to name de, fr or it, which FhirDownloader already did. A preliminary publication now exists in parallel under /static/sl/preliminary/fhir/foph-sl-preliminary-*, which oddb2xml does not use. The file prefix is its own constant now, so the next move touches one line.
+
+=== 3.0.31 / 14.07.2026
+* Bugfix (SLOPLUS / DEDUCTIBLE): every SL article claimed the raised 40% Selbstbehalt. The FHIR extractor mapped the BAG costShare percentage the wrong way round -- costShare 10 (the ordinary 10% Selbstbehalt, carried by ~97% of all packages) was turned into deductible "Y", which Builder#build_article emits as <SLOPLUS>1</SLOPLUS> (40%) and build_artikelstamm as <DEDUCTIBLE>40</DEDUCTIBLE>. costShare now maps 40 => "Y" (SLOPLUS 1 / DEDUCTIBLE 40), 10 => "N" (SLOPLUS 2 / DEDUCTIBLE 10), 20 => deductible20 "Y" (transitional rate, issue #81) -- the same semantics FlagSB carries in the legacy BAG Preparations.xml. The BAG publication was correct all along: the live feed has costShare 40 on 385 of 10'388 packages (222 preparations), 10 on the rest, and never mixes the two within one preparation.
+
 === 3.0.30 / 11.07.2026
 * New (-r/--rogger): prefer the German article names from the "Rogger Mediliste" (the name-conflict corrections Frau Rogger of Vitabyte/Zur Rose reports, task #OX-5985-1594). RoggerNames.load fetches the CSV export (GTIN,Mediname) of the shared Google Sheet directly at runtime -- sheet edits reach the feeds without any release step -- with a bundled fallback under data/ (a non-CSV response such as a Google sign-in page is rejected and falls back too). Builder#apply_rogger_name_overrides! replaces desc_de for every listed GTIN as the last step of the Refdata description cleanups, so the list wins over the issue-#112 fixes. German-only; FR/IT descriptions and all feeds without -r are unchanged. Initial list: 56 packs covering 13 conflict cases (METOJECT, ELTROXIN LF, CETIRIZIN/GABAPENTIN/ESOMEPRAZOL/EZETIMIB Spirig HC, ATOVAQUON PLUS, MOUNJARO KwikPen, MELATONIN, MONTELUKAST Mepha, FAMPYRA, RINVOQ).
 

data/README.md

@@ -294,12 +294,12 @@ We use the following files:
 * http://bag.e-mediat.net/SL2007.Web.External/File.axd?file=XMLPublications.zip
 * https://www.medregbm.admin.ch/Publikation/CreateExcelListBetriebs
 * https://www.medregbm.admin.ch/Publikation/CreateExcelListMedizinalPersons
-* http://zurrose.com/fileadmin/main/lib/download.php?file=/fileadmin/user_upload/downloads/ProduktUpdate/IGM11_mit_MwSt/Vollstamm/transfer.dat
+* https://www.zurrose.ch/sites/default/files/media/downloads/medi/produktupdate/igm11/de/transfer.dat (ZurRose IGM11 Vollstamm; mirrored by `scripts/get_transfer.sh`, which the gem then reads via `http://pillbox.oddb.org/TRANSFER.ZIP` or a local seed. The former `zurrose.com/fileadmin/main/lib/download.php?file=…` path now redirects to the marketing homepage.)
 * https://raw.githubusercontent.com/zdavatz/oddb2xml_files/master/NON-Pharma.xls
 * http://download.swissmedicinfo.ch/ (AipsDownload)
 * https://raw.githubusercontent.com/zdavatz/oddb2xml_files/master/LPPV.txt
 * https://raw.githubusercontent.com/zdavatz/cpp2sqlite/master/input/atc_codes_multi_lingual.txt
-* https://epl.bag.admin.ch/static/fhir/foph-sl-export-latest-{de,fr,it}.ndjson (FHIR NDJSON, used with `--fhir`)
+* https://epl.bag.admin.ch/static/sl/publication/fhir/foph-sl-publication-latest-{de,fr,it}.ndjson (FHIR NDJSON, used with `--fhir`)
 * https://id.gs1.ch/01/07612345000961 (GS1 Switzerland firstbase CSV — full barcode registry, used with `-b`/`--firstbase`)
 
 ## Indikationscode (BAG XXXXX.NN)
@@ -335,7 +335,7 @@ The same data is also exposed in-memory on each item and package as
 `item[:indication_codes]` (an array of `{code:, cud_id:, text:}`
 hashes) for downstream consumers that build their own emitters.
 
-Counted on the live FOPH feed (`foph-sl-export-latest-de.ndjson`,
+Counted on the live FOPH feed (`foph-sl-publication-latest-de.ndjson`,
 6'775 bundles): **539** products carry a total of **1'293** codes.
 
 Per BAG Rundschreiben vom 19. Februar 2026, the Indikationscode must
@@ -420,7 +420,7 @@ RegulatedAuthorization/65839
 ```
 
 The CUDs are identical in `id` across the three per-language NDJSON
-files (`foph-sl-export-latest-{de,fr,it}.ndjson`); only the
+files (`foph-sl-publication-latest-{de,fr,it}.ndjson`); only the
 `concept.text` differs. `FhirExtractor` resolves DE from the primary
 file and merges FR/IT in via the same CUD id. Coverage on the live
 feed went from 0 / 9'108 to 9'108 / 9'108 (100 %). See issue
@@ -520,6 +520,51 @@ a few times on such transient failures (`ODDB2XML_RETRIES`,
 `ODDB2XML_RETRY_DELAY`); a persistent per-host block still needs one of the
 mitigations above.
 
+## Deployment (the mediupdatexml.oddb.org download site)
+
+The `scripts/` directory automates the public download site
+<https://mediupdatexml.oddb.org>. These scripts are not part of the gem, but
+they are the reference setup for running oddb2xml unattended.
+
+* `get_transfer.sh` — mirrors the ZurRose IGM11 Vollstamm (`transfer.dat`) to
+  `/home/zdavatz/software/get_transfer/TRANSFER.ZIP` (cron 00:30), which the
+  nightly build seeds into its `downloads/` cache instead of fetching
+  `pillbox.oddb.org` — that host is only a mirror of this same job. Validates
+  the fixed-width record shape (98 bytes per record) before publishing, and
+  keeps the previous zip when the download fails, so a ZurRose outage cannot
+  abort the 01:00 build. Override the source with `TRANSFER_URL`.
+* `run_oddb2xml.sh` — nightly driver (cron 01:00). Downloads the sources once,
+  then builds the `-b` feed at price increments 45/50/55 plus `default` (no
+  increment) and the Artikelstamm (v6 + legacy v5), retrying transient download
+  failures (`ODDB2XML_RETRIES`, `ODDB2XML_RETRY_DELAY`).
+* `swissmedic_watch.sh` — every 30 min. While Swissmedic blocks this host it is
+  a silent no-op; the moment the site answers again it launches one build (at
+  most once per day).
+* `generate_index_html.sh` and `visitor_stats.py` — landing page with live
+  article counts and a visitor graph, refreshed hourly.
+* `setup_new_server.sh` — provisions a bare Debian host in one command:
+  packages, user permissions, the build/output directories,
+  `/etc/cron.d/mediupdatexml`, logrotate, then the Apache vhost and Let's
+  Encrypt certificate via `setup_mediupdatexml_web.sh`. Run as root; idempotent.
+* `setup_rust2xml.sh` — installs the Rust toolchain and builds
+  [rust2xml](https://github.com/zdavatz/rust2xml) for the 03:00 Artikelstamm
+  job. Run as the build user, not root.
+* `setup_aips2sqlite.sh` — installs the Java runtime, restores the
+  [aips2sqlite](https://github.com/zdavatz/aips2sqlite) checkout and creates
+  `jars/output`, the target of the `/aips2sqlite/` Apache alias, plus the 04:30
+  cron entry that regenerates the Fachinformationen. Run as root; idempotent.
+  Without it every `/aips2sqlite/` link answers 403, because Apache denies a
+  path that does not exist.
+* `transfer.sh` — optional scp hand-off of the output tree.
+
+One Debian-specific pitfall worth knowing when running oddb2xml from cron: with
+the system Ruby, `gem install oddb2xml` as an unprivileged user cannot write
+`/var/lib/gems` and falls back to a user install under
+`~/.local/share/gem/ruby/X.Y.0/bin`, which is on no default PATH — the build
+then fails with `oddb2xml: command not found`. `run_oddb2xml.sh` and
+`swissmedic_watch.sh` add that directory and `/usr/local/bin` to `PATH`
+themselves.
+
 ## Testing
 
 * Calling rake spec runs spec tests.

data/data/rogger_liste.csv

@@ -0,0 +1,57 @@
+GTIN,Mediname
+7680680690017,EZETIMIB ROSUVA Spirig HC Filmtab 10/10mg 30 Stk
+7680680690024,EZETIMIB ROSUVA Spirig HC Filmtab 10/10mg 100 Stk
+7680680690031,EZETIMIB ROSUVA Spirig HC Filmtab 10/20mg 30 Stk
+7680680690048,EZETIMIB ROSUVA Spirig HC Filmtab 10/20mg 100 Stk
+7680680690055,EZETIMIB ROSUVA Spirig HC Filmtab 10/5mg 30 Stk
+7680680690062,EZETIMIB ROSUVA Spirig HC Filmtab 10/5mg 100 Stk
+7680681050018,EZETIMIB ATORVAST Spirig HC Tab 10/10mg 30 Stk
+7680681050025,EZETIMIB ATORVAST Spirig HC Tab 10/10mg 90 Stk
+7680681050032,EZETIMIB ATORVAST Spirig HC Tab 10/20mg 30 Stk
+7680681050049,EZETIMIB ATORVAST Spirig HC Tab 10/20mg 90 Stk
+7680681050056,EZETIMIB ATORVAST Spirig HC Tab 10/40mg 30 Stk
+7680681050063,EZETIMIB ATORVAST Spirig HC Tab 10/40mg 90 Stk
+7680681050070,EZETIMIB ATORVAST Spirig HC Tab 10/80mg 30 Stk
+7680681050087,EZETIMIB ATORVAST Spirig HC Tab 10/80mg 90 Stk
+7680658280011,ESOMEPRAZOL Spirig HC Filmtabl 20 mg 14 Stk
+7680658280028,ESOMEPRAZOL Spirig HC Filmtabl 20 mg 30 Stk
+7680658280035,ESOMEPRAZOL Spirig HC Filmtabl 20 mg 60 Stk
+7680658280042,ESOMEPRAZOL Spirig HC Filmtabl 20 mg 100 Stk
+7680658280066,ESOMEPRAZOL Spirig HC Filmtabl 40 mg 14 Stk
+7680658280073,ESOMEPRAZOL Spirig HC Filmtabl 40 mg 30 Stk
+7680658280080,ESOMEPRAZOL Spirig HC Filmtabl 40 mg 60 Stk
+7680658280097,ESOMEPRAZOL Spirig HC Filmtabl 40 mg 100 Stk
+7680671390032,FAMPYRA Ret Tabl 10 mg 4x14 Stk
+7680665940014,GABAPENTIN Spirig HC Kaps 100 mg 50 Stk
+7680665940021,GABAPENTIN Spirig HC Kaps 100 mg 100 Stk
+7680665940038,GABAPENTIN Spirig HC Kaps 300 mg 50 Stk
+7680665940045,GABAPENTIN Spirig HC Kaps 300 mg 100 Stk
+7680665940052,GABAPENTIN Spirig HC Kaps 400 mg 50 Stk
+7680665940069,GABAPENTIN Spirig HC Kaps 400 mg 100 Stk
+7680656721011,METOJECT Autoinjektor 7.5 mg/0.15 ml Fertpen 1 Stk
+7680656721066,METOJECT Autoinjektor 10 mg/0.2 ml Fertpen 1 Stk
+7680656721110,METOJECT Autoinjektor 12.5 mg/0.25 ml Fertpen 1 Stk
+7680656721165,METOJECT Autoinjektor 15 mg/0.3 ml Fertpen 1 Stk
+7680656721219,METOJECT Autoinjektor 17.5 mg/0.35 ml Fertpen 1 Stk
+7680656721264,METOJECT Autoinjektor 20 mg/0.4 ml Fertpen 1 Stk
+7680656721318,METOJECT Autoinjektor 22.5 mg/0.45 ml Fertpen 1 Stk
+7680656721363,METOJECT Autoinjektor 25 mg/0.5 ml Fertpen 1 Stk
+7680298120210,ELTROXIN LF Tabl 50 µg 100 Stk
+7680298120012,ELTROXIN LF Tabl 100 µg 100 Stk
+7680625680042,CETIRIZIN Spirig HC Filmtabl 10 mg 10 Stk
+7680625680073,CETIRIZIN Spirig HC Filmtabl 10 mg 30 Stk
+7680625680080,CETIRIZIN Spirig HC Filmtabl 10 mg 50 Stk
+7680652800017,ATOVAQUON PLUS Spirig HC Filmtabl 250 mg / 100 mg 12 Stk
+7680652800024,ATOVAQUON PLUS Spirig HC Filmtabl 250 mg / 100 mg 24 Stk
+7680696960012,MOUNJARO KwikPen Inj Lös 2.5 mg/0.6 ml 1 Stk
+7680696960029,MOUNJARO KwikPen Inj Lös 5 mg/0.6 ml 1 Stk
+7680696960036,MOUNJARO KwikPen Inj Lös 7.5 mg/0.6 ml 1 Stk
+7680696960043,MOUNJARO KwikPen Inj Lös 10 mg/0.6 ml 1 Stk
+7680696960050,MOUNJARO KwikPen Inj Lös 12.5 mg/0.6 ml 1 Stk
+7680696960067,MOUNJARO KwikPen Inj Lös 15 mg/0.6 ml 1 Stk
+7680689680019,MELATONIN Spirig HC Ret Tabl 2 mg 21 Stk
+7680689680026,MELATONIN Spirig HC Ret Tabl 2 mg 100 Stk
+7680623620019,MONTELUKAST Mepha Lactab 10 mg 28 Stk
+7680672570013,RINVOQ Ret Tabl 15 mg 28 Stk
+7680672570037,RINVOQ Ret Tabl 30 mg 28 Stk
+7680672570044,RINVOQ Ret Tabl 45 mg 28 Stk

data/lib/oddb2xml/fhir_support.rb

@@ -14,7 +14,17 @@ module Oddb2xml
     include DownloadMethod
 
     BASE_URL = "https://epl.bag.admin.ch"
-    STATIC_FHIR_PATH = "/static/fhir"
+    # BAG moved the export on 24.08.2026 and announced it the next evening:
+    # /static/fhir/foph-sl-export-* became
+    # /static/sl/publication/fhir/foph-sl-publication-*.
+    #
+    # The old -latest- alias answers 404 while the old dated snapshots stay in
+    # place, so this fails by downloading nothing rather than by crashing.
+    # There is no language-less default any more; per BAG each URL must name
+    # de, fr or it. A preliminary publication now exists in parallel under
+    # /static/sl/preliminary/fhir/foph-sl-preliminary-*, which this does not use.
+    STATIC_FHIR_PATH = "/static/sl/publication/fhir"
+    FILE_PREFIX = "foph-sl-publication"
     LANGUAGES = %w[de fr it].freeze
 
     def initialize(options = {})
@@ -32,7 +42,7 @@ module Oddb2xml
       else
         files = {}
         LANGUAGES.each do |lang|
-          url = "#{BASE_URL}#{STATIC_FHIR_PATH}/foph-sl-export-latest-#{lang}.ndjson"
+          url = "#{BASE_URL}#{STATIC_FHIR_PATH}/#{FILE_PREFIX}-latest-#{lang}.ndjson"
           path = download_one(url)
           files[lang] = path if path
         end
@@ -832,16 +842,19 @@ module Oddb2xml
             end
             item[:packages][ean13][:limitation_points] = ""
             
-            # Map cost_share to deductible flags
+            # Map the BAG costShare percentage to the deductible flags. "Y"
+            # means *erhoehter* Selbstbehalt (40%, SLOPLUS 1 / DEDUCTIBLE 40),
+            # "N" the ordinary one (10%, SLOPLUS 2 / DEDUCTIBLE 10) -- the same
+            # semantics FlagSB carries in the legacy BAG Preparations.xml.
             if pac.CostShare
               case pac.CostShare
-              when 10
+              when 40
                 item[:deductible] = "Y"
               when 20
+                # Transitional 20% rate, see issue #81. Kept for old feeds.
                 item[:deductible20] = "Y"
-              when 40
-                # New value - might need new field or special handling
-                item[:deductible] = "Y"  # Fallback to standard deductible
+              else # 10 (the vast majority) and anything unexpected
+                item[:deductible] = "N"
               end
             end
 

data/lib/oddb2xml/rogger_names.rb

@@ -0,0 +1,86 @@
+require "csv"
+require "oddb2xml/downloader"
+
+module Oddb2xml
+  # Preferred German article names from the "Rogger Mediliste" — the
+  # name-conflict list maintained by Frau Rogger (Vitabyte/Zur Rose, task
+  # #OX-5985-1594). The source of truth is the shared Google Sheet
+  # "Rogger Mediliste" (GTIN,Mediname); RoggerDownloader fetches its CSV
+  # export directly, so sheet edits reach the feeds without any release step.
+  # A bundled copy under data/ serves as offline fallback (refresh it at
+  # release time when the sheet changed). A response that is not the expected
+  # CSV (e.g. a Google sign-in page when the sheet is not link-shared) is
+  # rejected and the fallback engages.
+  #
+  # Activated with -r/--rogger: for every GTIN on the list the German
+  # description coming from Refdata is replaced by the list's Mediname. The
+  # list is German-only, so FR/IT descriptions are left untouched. Applied as
+  # the last step of Builder#apply_refdata_description_cleanups!, so it sees
+  # (and wins over) the issue-#112 Refdata cleanups.
+  module RoggerNames
+    DATA_DIR = File.expand_path(File.join(__dir__, "..", "..", "data"))
+
+    module_function
+
+    # Returns a Hash keyed by the 13-digit GTIN (String):
+    #   "7680672570037" => "RINVOQ Ret Tabl 30 mg 28 Stk"
+    # Returns {} if the data cannot be obtained (never raises — the rest of
+    # the build must proceed).
+    def load(options = {})
+      map = parse(source(options))
+      Oddb2xml.log "RoggerNames: #{map.size} preferred names loaded"
+      map
+    rescue => error
+      Oddb2xml.log "RoggerNames: disabled (#{error.class}: #{error.message})"
+      {}
+    end
+
+    # Download the Google Sheet CSV export; fall back to the bundled copy
+    # under data/ when the download is unavailable (offline, allow-list proxy)
+    # or does not look like the expected CSV (sheet not link-shared, Google
+    # error/sign-in page).
+    def source(options)
+      content = nil
+      begin
+        content = RoggerDownloader.new(options).download
+      rescue => error
+        Oddb2xml.log "RoggerNames: download of rogger_liste.csv failed (#{error.class}: #{error.message})"
+      end
+      unless rogger_csv?(content)
+        bundled = File.join(DATA_DIR, "rogger_liste.csv")
+        if File.exist?(bundled)
+          Oddb2xml.log "RoggerNames: using bundled rogger_liste.csv"
+          content = File.read(bundled, encoding: "UTF-8")
+        end
+      end
+      content
+    end
+
+    # True when the content is the expected sheet export: a CSV whose header
+    # row carries the GTIN and Mediname columns (rejects empty bodies and
+    # HTML sign-in/error pages).
+    def rogger_csv?(content)
+      return false if content.nil?
+      header = content.to_s.dup.force_encoding(Encoding::UTF_8)
+        .sub(/\A\xEF\xBB\xBF/, "").lines.first.to_s
+      /GTIN/i.match?(header) && /Mediname/i.match?(header)
+    end
+
+    def parse(csv_string)
+      map = {}
+      return map if csv_string.nil?
+      # The list carries non-ASCII (e.g. "µg"); guard against a US-ASCII
+      # default external encoding.
+      csv_string = csv_string.dup.force_encoding(Encoding::UTF_8) unless csv_string.encoding == Encoding::UTF_8
+      return map if csv_string.strip.empty?
+      CSV.parse(csv_string, headers: true) do |row|
+        gtin = row["GTIN"].to_s.strip.rjust(13, "0")
+        name = row["Mediname"].to_s.strip
+        next unless /\A\d{13}\z/.match?(gtin)
+        next if name.empty?
+        map[gtin] = name
+      end
+      map
+    end
+  end
+end

data/lib/oddb2xml/version.rb

@@ -1,3 +1,3 @@
 module Oddb2xml
-  VERSION = "3.0.30"
+  VERSION = "3.0.32"
 end

data/scripts/get_transfer.sh

@@ -0,0 +1,75 @@
+#!/usr/bin/env bash
+#
+# get_transfer — mirror the ZurRose IGM11 "Vollstamm" article master locally.
+#
+# ZurRose publishes the full article master as a fixed-width transfer.dat. The
+# nightly build (run_oddb2xml.sh, 01:00) seeds the zip produced here into its
+# downloads/ cache, so oddb2xml reads ZurRose from this host instead of
+# fetching http://pillbox.oddb.org/TRANSFER.ZIP — see seed_downloads() there.
+#
+# History worth keeping: this script used to live ONLY in
+# /home/zdavatz/software/get_transfer on the old server, so when that host was
+# deleted (2026-07-28) the mirror went with it, together with its download URL.
+# It now lives in the repo and $DATA_DIR/get_transfer.sh is a symlink to it, so
+# a rebuilt server gets it back from the checkout (setup_new_server.sh).
+#
+# The download URL moved too: the old
+# zurrose.com/fileadmin/main/lib/download.php?file=... path 301s to the
+# marketing homepage since roughly mid-July 2026. The current one is below.
+#
+# ZurRose refreshes the master roughly every 14 days, so most runs legitimately
+# fetch a file identical to yesterday's; that is not an error.
+#
+# Configurable via environment:
+#   TRANSFER_URL   source URL       (default: the ZurRose IGM11 de/transfer.dat)
+#   DATA_DIR       where TRANSFER.ZIP lands (default /home/zdavatz/software/get_transfer)
+#   MIN_LINES      reject a file with fewer records  (default 100000)
+#
+set -euo pipefail
+
+TRANSFER_URL="${TRANSFER_URL:-https://www.zurrose.ch/sites/default/files/media/downloads/medi/produktupdate/igm11/de/transfer.dat}"
+DATA_DIR="${DATA_DIR:-/home/zdavatz/software/get_transfer}"
+MIN_LINES="${MIN_LINES:-100000}"
+
+ZIP="$DATA_DIR/TRANSFER.ZIP"
+
+log() { printf '%s %s\n' "$(date '+%Y-%m-%d %H:%M:%S')" "$*"; }
+die() { log "ERROR: $*"; exit 1; }
+
+mkdir -p "$DATA_DIR"
+
+# Work in a scratch dir so a failed or truncated download can never replace a
+# good TRANSFER.ZIP. The build reads that zip unconditionally at 01:00; half a
+# file there would abort the whole nightly run at oddb2xml's unzip step.
+WORK="$(mktemp -d "${TMPDIR:-/tmp}/get_transfer.XXXXXX")"
+trap 'rm -rf "$WORK"' EXIT
+
+log "Downloading $TRANSFER_URL"
+curl -fsSL --max-time 900 --retry 3 --retry-delay 30 \
+     -o "$WORK/transfer.dat" "$TRANSFER_URL" \
+  || die "download failed - keeping $( [[ -s $ZIP ]] && echo "the previous TRANSFER.ZIP" || echo "nothing, there is no previous TRANSFER.ZIP" )"
+
+# Validate before publishing. transfer.dat is a fixed-width IGM11 file: every
+# record is exactly 98 bytes. A login page, an error page or a truncated
+# transfer would sail through a plain size check but fails this one.
+lines=$(wc -l < "$WORK/transfer.dat")
+(( lines >= MIN_LINES )) \
+  || die "only $lines records (< $MIN_LINES) - looks like an error page, not the Vollstamm"
+badlen=$(awk 'length != 98 { n++ } END { print n+0 }' "$WORK/transfer.dat")
+(( badlen == 0 )) \
+  || die "$badlen of $lines records are not 98 chars - not an IGM11 fixed-width file"
+
+# The zip entry MUST be named transfer.dat (lowercase): oddb2xml unzips it into
+# downloads/ and then opens downloads/transfer.dat by that exact name
+# (ZurroseDownloader, lib/oddb2xml/downloader.rb).
+( cd "$WORK" && zip -q TRANSFER.ZIP transfer.dat )
+
+if cmp -s "$WORK/transfer.dat" <(unzip -p "$ZIP" transfer.dat 2>/dev/null); then
+  log "Unchanged since the last run ($lines records) - refreshing the mirror anyway"
+fi
+
+# Atomic publish: same filesystem, so mv is a rename. A reader either sees the
+# old zip or the new one, never a partial.
+mv "$WORK/TRANSFER.ZIP" "$ZIP.new" && mv "$ZIP.new" "$ZIP"
+
+log "Wrote $ZIP ($(stat -c %s "$ZIP") bytes, $lines records)"

data/scripts/run_oddb2xml.sh

@@ -70,9 +70,17 @@ run_with_retry() {
 }
 
 # 1. Install / update the published gem unless told otherwise.
+# Under Debian's system Ruby an unprivileged `gem install` cannot write
+# /var/lib/gems and silently falls back to a user install, whose bin dir is on
+# nobody's PATH — so the freshly installed gem is unreachable and the build dies
+# with exit 127. Put both candidate dirs on PATH, user install first, so
+# $ODDB2XML_BIN resolves to whatever this step just installed.
 if [[ "${SKIP_GEM_INSTALL:-0}" != "1" ]]; then
   log "Installing oddb2xml gem"
   gem install oddb2xml
+  GEM_USER_BIN="$(ruby -e 'print Gem.user_dir' 2>/dev/null)/bin"
+  [[ -d "$GEM_USER_BIN" ]] && export PATH="$GEM_USER_BIN:$PATH"
+  case ":$PATH:" in */usr/local/bin:*) ;; *) export PATH="/usr/local/bin:$PATH" ;; esac
 fi
 
 # 2. Fresh working dir (keeps a shared downloads/ cache across increments).

data/scripts/setup_aips2sqlite.sh

@@ -0,0 +1,143 @@
+#!/usr/bin/env bash
+#
+# setup_aips2sqlite.sh — provision the aips2sqlite half of mediupdatexml.oddb.org.
+#
+# The landing page links a whole section of aips2sqlite output (Fachinformationen
+# as XML, the AmiKo databases, the Swissmedic-sequences CSV), served by Apache
+# through the alias
+#
+#     /aips2sqlite/  ->  $AIPS_DIR/jars/output
+#
+# written by setup_mediupdatexml_web.sh. That alias is only half the story: the
+# directory it points at is produced by a Java job that has to run on this host.
+# On the rebuilt server (2026-07-28) neither the JRE nor the output directory
+# existed, so every /aips2sqlite/ link answered 403 — Apache denies a path that
+# does not exist, because the granting <Directory> block never matches.
+#
+# This script installs the runtime, restores the checkout, creates the output
+# directory and schedules the nightly regeneration.
+#
+# Run with:  sudo scripts/setup_aips2sqlite.sh
+#
+# Idempotent — safe to re-run. It does NOT generate the data (that takes ~1 h
+# and belongs to the build user); run scripts/generate_aips_fi afterwards, or
+# wait for the cron entry it installs.
+#
+# Configurable via environment:
+#   RUN_USER    build/cron user       (default zdavatz)
+#   AIPS_DIR    checkout location     (default /home/<RUN_USER>/software/aips2sqlite)
+#   AIPS_REPO   clone URL             (default https://github.com/zdavatz/aips2sqlite)
+#   OUT_DIR     published root, only used to locate the log dir
+#   SKIP_CRON   set to 1 to leave /etc/cron.d/mediupdatexml alone
+#
+set -euo pipefail
+
+RUN_USER="${RUN_USER:-zdavatz}"
+RUN_HOME="$(getent passwd "$RUN_USER" | cut -d: -f6)"
+AIPS_DIR="${AIPS_DIR:-${RUN_HOME}/software/aips2sqlite}"
+AIPS_REPO="${AIPS_REPO:-https://github.com/zdavatz/aips2sqlite}"
+OUT_DIR="${OUT_DIR:-${RUN_HOME}/oddb2xml}"
+STATE_DIR="${OUT_DIR%/}-state"
+CRON_FILE=/etc/cron.d/mediupdatexml
+
+step() { printf '\n\033[1m==> %s\033[0m\n' "$*"; }
+
+if [[ $EUID -ne 0 ]]; then
+  echo "Please run as root:  sudo $0" >&2
+  exit 1
+fi
+if [[ -z "$RUN_HOME" ]]; then
+  echo "User $RUN_USER does not exist" >&2
+  exit 1
+fi
+
+# --- 1. Java runtime ---------------------------------------------------------
+# aips2sqlite ships a prebuilt fat jar (jars/aips2sqlite.jar), so a JRE is
+# enough — no JDK and no Gradle unless the jar itself is rebuilt. README asks
+# for Java 21+; Debian 13 has both 21 and 25, pin 21 to match what the jar was
+# built and tested against.
+#
+# The font libraries are NOT optional, even though Debian lists them as mere
+# Recommends of the headless JRE: every Fachinfo carries a rendered EAN13
+# barcode (RealExpertInfo.updateSectionPackungen -> BarCode.encode -> barcode4j),
+# and drawing its digits loads libfontmanager.so. Installing the JRE with
+# --no-install-recommends therefore gets through the whole download phase and
+# then dies on the first medicine with
+#   UnsatisfiedLinkError: libfontmanager.so: libharfbuzz.so.0: cannot open ...
+# so they are listed explicitly here rather than left to apt's recommends
+# handling. fonts-dejavu-core gives fontconfig an actual font to find.
+step "Installing the Java runtime (openjdk-21-jre-headless + font stack)"
+export DEBIAN_FRONTEND=noninteractive
+apt-get update -qq
+apt-get install -y --no-install-recommends \
+  openjdk-21-jre-headless \
+  libfreetype6 libharfbuzz0b libfontconfig1 fonts-dejavu-core
+java -version
+
+# --- 2. Checkout -------------------------------------------------------------
+step "Ensuring the aips2sqlite checkout at $AIPS_DIR"
+if [[ -d "$AIPS_DIR/.git" ]]; then
+  echo "already present — leaving it alone (generate_aips_fi does its own git pull)"
+else
+  install -d -o "$RUN_USER" -g "$RUN_USER" "$(dirname "$AIPS_DIR")"
+  sudo -u "$RUN_USER" git clone "$AIPS_REPO" "$AIPS_DIR"
+fi
+
+# --- 3. Output + download directories ----------------------------------------
+# jars/output is the Apache alias target: it must exist even while empty, or
+# /aips2sqlite/ answers 403 instead of an (empty) listing.
+step "Creating $AIPS_DIR/jars/{output,downloads}"
+for d in "$AIPS_DIR/jars/output" "$AIPS_DIR/jars/downloads"; do
+  install -d -o "$RUN_USER" -g "$RUN_USER" -m 755 "$d"
+done
+
+# --- 4. Nightly regeneration -------------------------------------------------
+# The schedule lives in the one cron file setup_new_server.sh writes; append
+# only when that file exists and does not already carry the entry, so the two
+# scripts can be run in either order without fighting over it.
+if [[ "${SKIP_CRON:-0}" != "1" ]]; then
+  step "Scheduling the nightly Fachinfo generation (04:30)"
+  install -d -o "$RUN_USER" -g "$RUN_USER" "$STATE_DIR"
+  if [[ ! -f "$CRON_FILE" ]]; then
+    echo "!! $CRON_FILE does not exist yet — run setup_new_server.sh first,"
+    echo "   it writes the full schedule including this entry."
+  elif grep -q 'generate_aips_fi' "$CRON_FILE"; then
+    echo "entry already present in $CRON_FILE"
+  else
+    cat >> "$CRON_FILE" <<EOF
+
+# 04:30  aips2sqlite Fachinfo XMLs + AmiKo DBs + Swissmedic sequences,
+#        published under /aips2sqlite/.
+30 4 * * *   $RUN_USER  [ -x $AIPS_DIR/scripts/generate_aips_fi ] && $AIPS_DIR/scripts/generate_aips_fi >> $STATE_DIR/generate_aips_fi.log 2>&1
+EOF
+    echo "appended to $CRON_FILE"
+  fi
+fi
+
+cat <<EOF
+
+==> aips2sqlite provisioning done.
+
+The alias /aips2sqlite/ -> $AIPS_DIR/jars/output now resolves; it stays empty
+until the generator has run once. As $RUN_USER (takes ~1 h, ~8 GB heap):
+
+    $AIPS_DIR/scripts/generate_aips_fi
+
+Optional: export REFDATA_API_KEY (developer.refdata.ch) to also fetch the
+Refdata Partner GLN data. It is only consumed by the Takeda partner export,
+so the Fachinfo/sequences output is complete without it — the run just logs
+one download exception.
+
+Known upstream gap: BAG's resource index
+https://epl.bag.admin.ch/api/sl/public/resources/current currently reports
+"fhir": {"fileUrl": null}, so aips2sqlite's FHIR download resolves to
+/static/null and the run parses 0 preparations (no SL flags, no prices). The
+export itself is alive at the fixed path oddb2xml uses:
+
+    https://epl.bag.admin.ch/static/sl/publication/fhir/foph-sl-publication-latest-de.ndjson
+
+Until the jar learns that fallback, seed it before the run:
+
+    cp <oddb2xml downloads>/foph-sl-publication-latest-de.ndjson \\
+       $AIPS_DIR/jars/downloads/fhir-sl.ndjson
+EOF

data/scripts/setup_new_server.sh

@@ -0,0 +1,183 @@
+#!/usr/bin/env bash
+#
+# setup_new_server.sh — bring a bare Debian host up to the point where it can
+# run the mediupdatexml.oddb.org download site and its nightly builds.
+#
+# Written after HIN deleted the previous server (2026-07-28): everything that
+# had accumulated by hand on that box is captured here, so a rebuild is one
+# command instead of a day of archaeology.
+#
+# Run with:  sudo scripts/setup_new_server.sh
+#
+# Idempotent — safe to re-run. It installs packages, creates the build/output
+# directories, installs the oddb2xml gem, writes /etc/cron.d/mediupdatexml and
+# a logrotate rule, and finally hands over to setup_mediupdatexml_web.sh for the
+# Apache vhost + Let's Encrypt certificate.
+#
+# What it deliberately does NOT do (run these as the build user afterwards):
+#   scripts/setup_rust2xml.sh   Rust toolchain for the rust2xml artikelstamm job
+#
+# Configurable via environment:
+#   RUN_USER   build/cron user            (default zdavatz)
+#   OUT_DIR    published download root    (default /home/<RUN_USER>/oddb2xml)
+#   SKIP_WEB   set to 1 to skip the Apache/certbot step
+#   SKIP_GEM   set to 1 to skip `gem install oddb2xml`
+#   SKIP_AIPS  set to 1 to skip the aips2sqlite provisioning
+#
+set -euo pipefail
+
+RUN_USER="${RUN_USER:-zdavatz}"
+RUN_HOME="$(getent passwd "$RUN_USER" | cut -d: -f6)"
+OUT_DIR="${OUT_DIR:-${RUN_HOME}/oddb2xml}"
+BUILD_DIR="${OUT_DIR%/}-build"
+STATE_DIR="${OUT_DIR%/}-state"
+WATCH_DIR="${OUT_DIR%/}-watch"
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+REPO_DIR="$(dirname "$SCRIPT_DIR")"
+RUST2XML_DIR="${RUST2XML_DIR:-${RUN_HOME}/software/rust2xml}"
+GET_TRANSFER_DIR="${GET_TRANSFER_DIR:-${RUN_HOME}/software/get_transfer}"
+AIPS_DIR="${AIPS_DIR:-${RUN_HOME}/software/aips2sqlite}"
+
+step() { printf '\n\033[1m==> %s\033[0m\n' "$*"; }
+
+if [[ $EUID -ne 0 ]]; then
+  echo "Please run as root:  sudo $0" >&2
+  exit 1
+fi
+if [[ -z "$RUN_HOME" ]]; then
+  echo "User $RUN_USER does not exist" >&2
+  exit 1
+fi
+
+# --- 1. System packages ------------------------------------------------------
+# ruby-full     Debian 13 ships Ruby 3.3, matching the repo's .ruby-version --
+#               no rbenv needed (the old box carried an rbenv 3.4.5 install).
+# unzip/zip     run_oddb2xml.sh unpacks the build zip; ZurroseDownloader shells
+#               out to `unzip` for transfer.zip, and to `iconv` (libc-bin).
+# cron          not installed by default on a minimal Debian 13 image.
+# poppler-utils pdftotext, for tools/generate_bag_sl_group_prices.rb.
+# lib*-dev      native build fallbacks for nokogiri/ffi when no precompiled gem
+#               matches the platform.
+step "Installing system packages"
+export DEBIAN_FRONTEND=noninteractive
+apt-get update -qq
+apt-get install -y --no-install-recommends \
+  apache2 certbot python3-certbot-apache \
+  ruby-full ruby-dev build-essential \
+  cron unzip zip curl wget git ca-certificates \
+  poppler-utils libxml2-dev libxslt1-dev zlib1g-dev libyaml-dev pkg-config
+systemctl enable --now cron >/dev/null 2>&1 || true
+
+# --- 2. User, groups, permissions -------------------------------------------
+# "adm" lets visitor_stats.py read /var/log/apache2/*access.log for the graph.
+step "Granting $RUN_USER read access to the Apache logs (group adm)"
+usermod -aG adm "$RUN_USER"
+
+# www-data must traverse (not list) the home dir to reach $OUT_DIR.
+step "Setting $RUN_HOME to 711 so Apache can traverse it"
+chmod 711 "$RUN_HOME"
+
+# --- 3. The gem --------------------------------------------------------------
+if [[ "${SKIP_GEM:-0}" != "1" ]]; then
+  step "Installing the oddb2xml gem (system-wide, binstub in /usr/local/bin)"
+  gem install oddb2xml --no-document
+  oddb2xml --version 2>/dev/null || true
+fi
+
+# --- 4. Directory layout -----------------------------------------------------
+# OUT_DIR    published, served by Apache (per-increment subdirs + landing page)
+# BUILD_DIR  shared downloads/ cache; wiped at the start of every build
+# STATE_DIR  last-good firstbase.csv + cron logs; must survive the wipe
+# WATCH_DIR  swissmedic_watch.sh stamps and log
+# GET_TRANSFER_DIR  local ZurRose mirror (TRANSFER.ZIP), seeded into the build
+step "Creating $OUT_DIR, $BUILD_DIR, $STATE_DIR, $WATCH_DIR, $GET_TRANSFER_DIR"
+for d in "$OUT_DIR" "$BUILD_DIR" "$STATE_DIR" "$WATCH_DIR" "$GET_TRANSFER_DIR"; do
+  mkdir -p "$d"
+  chown "$RUN_USER:$RUN_USER" "$d"
+done
+
+# The ZurRose mirror script used to exist only in $GET_TRANSFER_DIR on the old
+# server, so deleting that host lost both the script and its download URL, and
+# the cron line below — guarded by [ -x ] — went quietly dead. Symlink the
+# tracked copy instead, so a rebuild restores the mirror from the checkout.
+ln -sfn "$SCRIPT_DIR/get_transfer.sh" "$GET_TRANSFER_DIR/get_transfer.sh"
+chown -h "$RUN_USER:$RUN_USER" "$GET_TRANSFER_DIR/get_transfer.sh"
+
+# --- 5. Cron -----------------------------------------------------------------
+# One /etc/cron.d file instead of the old scattered per-user crontab, so the
+# whole schedule is visible (and re-creatable) in one place.
+step "Writing /etc/cron.d/mediupdatexml"
+cat > /etc/cron.d/mediupdatexml <<EOF
+# mediupdatexml.oddb.org — nightly feed builds and download-site upkeep.
+# Managed by oddb2xml/scripts/setup_new_server.sh; edit there, not here.
+SHELL=/bin/bash
+PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
+MAILTO=$RUN_USER
+
+# 00:30  Mirror the ZurRose transfer.dat locally (seeds the build's downloads/).
+30 0 * * *   $RUN_USER  [ -x $GET_TRANSFER_DIR/get_transfer.sh ] && $GET_TRANSFER_DIR/get_transfer.sh >> $STATE_DIR/get_transfer.log 2>&1
+
+# 01:00  Nightly oddb2xml build: default + artikelstamm (v6/v5) + 45/50/55.
+0 1 * * *    $RUN_USER  $SCRIPT_DIR/run_oddb2xml.sh >> $STATE_DIR/run_oddb2xml.log 2>&1
+
+# 03:00  rust2xml's own Artikelstamm, published under /artikelstamm/rust2xml/.
+0 3 * * *    $RUN_USER  [ -x $RUST2XML_DIR/scripts/run_artikelstamm.sh ] && $RUST2XML_DIR/scripts/run_artikelstamm.sh >> $STATE_DIR/run_artikelstamm.log 2>&1
+
+# 04:30  aips2sqlite Fachinfo XMLs + AmiKo DBs + Swissmedic sequences,
+#        published under /aips2sqlite/.
+30 4 * * *   $RUN_USER  [ -x $AIPS_DIR/scripts/generate_aips_fi ] && $AIPS_DIR/scripts/generate_aips_fi >> $STATE_DIR/generate_aips_fi.log 2>&1
+
+# hourly  Refresh the landing page (live counts + visitor graph).
+5 * * * *    $RUN_USER  $SCRIPT_DIR/generate_index_html.sh $OUT_DIR >> $STATE_DIR/index_html.log 2>&1
+
+# every 30 min  Rebuild automatically once Swissmedic answers again after a block.
+*/30 * * * * $RUN_USER  $SCRIPT_DIR/swissmedic_watch.sh
+EOF
+chmod 644 /etc/cron.d/mediupdatexml
+
+step "Writing /etc/logrotate.d/mediupdatexml"
+cat > /etc/logrotate.d/mediupdatexml <<EOF
+$STATE_DIR/*.log $WATCH_DIR/*.log {
+    weekly
+    rotate 8
+    compress
+    delaycompress
+    missingok
+    notifempty
+    su $RUN_USER $RUN_USER
+    create 0644 $RUN_USER $RUN_USER
+}
+EOF
+chmod 644 /etc/logrotate.d/mediupdatexml
+
+# --- 6. aips2sqlite (JRE, checkout, /aips2sqlite alias target) ----------------
+# The landing page links a whole section of aips2sqlite output; without this the
+# alias target is missing and every /aips2sqlite/ link answers 403.
+if [[ "${SKIP_AIPS:-0}" != "1" ]]; then
+  step "Provisioning aips2sqlite (setup_aips2sqlite.sh)"
+  SKIP_CRON=1 "$SCRIPT_DIR/setup_aips2sqlite.sh"   # cron entry already written above
+fi
+
+# --- 7. Apache vhost + HTTPS -------------------------------------------------
+if [[ "${SKIP_WEB:-0}" != "1" ]]; then
+  step "Provisioning the Apache vhost (setup_mediupdatexml_web.sh)"
+  "$SCRIPT_DIR/setup_mediupdatexml_web.sh"
+fi
+
+cat <<EOF
+
+==> Host provisioning done.
+
+Next steps (as $RUN_USER, not root):
+  1. ZurRose mirror (needed by step 3):  $GET_TRANSFER_DIR/get_transfer.sh
+  2. Rust toolchain for the 03:00 job:   $SCRIPT_DIR/setup_rust2xml.sh
+  3. First build (takes ~1-2 h):         $SCRIPT_DIR/run_oddb2xml.sh
+  4. First Fachinfo run (takes ~1 h):    $AIPS_DIR/scripts/generate_aips_fi
+  5. Landing page:                       $SCRIPT_DIR/generate_index_html.sh $OUT_DIR
+
+Step 1 before step 3: without $GET_TRANSFER_DIR/TRANSFER.ZIP the build falls
+back to http://pillbox.oddb.org/TRANSFER.ZIP, which is itself only a mirror of
+this job and has gone stale (and refused connections at 01:00) before.
+
+Check the schedule with:  cat /etc/cron.d/mediupdatexml
+EOF

data/scripts/setup_rust2xml.sh

@@ -0,0 +1,50 @@
+#!/usr/bin/env bash
+#
+# setup_rust2xml.sh — install the Rust toolchain and build rust2xml, so the
+# 03:00 cron job (rust2xml/scripts/run_artikelstamm.sh) can publish its
+# Artikelstamm under https://mediupdatexml.oddb.org/artikelstamm/rust2xml/.
+#
+# Run as the build user (NOT root):  scripts/setup_rust2xml.sh
+#
+# Idempotent: re-running updates the toolchain and rebuilds if sources changed.
+# rustup is used rather than Debian's rustc because run_artikelstamm.sh sources
+# $HOME/.cargo/env to find cargo under cron's minimal PATH.
+#
+# Configurable via environment:
+#   RUST2XML_DIR  rust2xml checkout  (default $HOME/software/rust2xml)
+#
+set -euo pipefail
+
+RUST2XML_DIR="${RUST2XML_DIR:-$HOME/software/rust2xml}"
+
+step() { printf '\n\033[1m==> %s\033[0m\n' "$*"; }
+
+if [[ $EUID -eq 0 ]]; then
+  echo "Run this as the build user, not root (rustup installs into \$HOME)." >&2
+  exit 1
+fi
+if [[ ! -d "$RUST2XML_DIR" ]]; then
+  echo "rust2xml checkout not found at $RUST2XML_DIR" >&2
+  echo "  git clone https://github.com/zdavatz/rust2xml $RUST2XML_DIR" >&2
+  exit 1
+fi
+
+if [[ ! -x "$HOME/.cargo/bin/cargo" ]]; then
+  step "Installing the Rust toolchain (rustup, stable)"
+  curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --no-modify-path
+else
+  step "Updating the Rust toolchain"
+  "$HOME/.cargo/bin/rustup" update stable
+fi
+
+# shellcheck disable=SC1091
+source "$HOME/.cargo/env"
+cargo --version
+
+step "Building rust2xml (release)"
+cd "$RUST2XML_DIR"
+cargo build --release --bin rust2xml
+
+step "Done"
+echo "Binary: $RUST2XML_DIR/target/release/rust2xml"
+echo "The 03:00 cron entry in /etc/cron.d/mediupdatexml picks it up from here."

data/scripts/swissmedic_watch.sh

@@ -27,10 +27,20 @@ STATE_DIR="${STATE_DIR:-${OUT_DIR%/}-watch}"
 ARTICLE_XML="$OUT_DIR/default/oddb_article.xml"
 CANARY_URL="https://www.swissmedic.ch/swissmedic/de/home/services/listen_neu.html"
 
-# Match the nightly cron's rbenv environment (the repo's .ruby-version points at
-# a Ruby that isn't installed here; cron pins RBENV_VERSION instead).
-export RBENV_VERSION="${RBENV_VERSION:-3.4.5}"
-export PATH="/home/zdavatz/.rbenv/shims:/usr/bin:/bin"
+# Cron hands us a minimal PATH, so spell out where the oddb2xml binstub can be,
+# covering every way this host might have installed it: Debian's system Ruby
+# puts a root `gem install` in /usr/local/bin, an unprivileged one in
+# ~/.local/share/gem/ruby/X.Y.0/bin, and the pre-2026-07 box used rbenv shims.
+# Getting this wrong is silent until build time — the watcher launched a run on
+# the rebuilt server with an rbenv-only PATH and it died with exit 127.
+export PATH="/usr/local/bin:/usr/bin:/bin"
+HOME_DIR="${HOME:-/home/zdavatz}"
+if [[ -d "$HOME_DIR/.rbenv/shims" ]]; then
+  export RBENV_VERSION="${RBENV_VERSION:-3.4.5}"
+  export PATH="$HOME_DIR/.rbenv/shims:$PATH"
+fi
+GEM_USER_BIN="$(ruby -e 'print Gem.user_dir' 2>/dev/null)/bin"
+[[ -d "$GEM_USER_BIN" ]] && export PATH="$GEM_USER_BIN:$PATH"
 
 mkdir -p "$STATE_DIR"
 LOG="$STATE_DIR/swissmedic_watch.log"

data/spec/fhir_spec.rb

@@ -102,6 +102,53 @@ describe "FHIR Indikationscode support" do
     end
   end
 
+  describe Oddb2xml::FhirExtractor, "Selbstbehalt (costShare)" do
+    # deductible "N" => SLOPLUS 2 / DEDUCTIBLE 10 (ordinary 10% Selbstbehalt),
+    # "Y" => SLOPLUS 1 / DEDUCTIBLE 40 (erhoehter Selbstbehalt).
+    def extract_with_cost_share(percent)
+      bundle = JSON.parse(File.read(cyramza_fixture))
+      bundle["entry"].each do |entry|
+        res = entry["resource"]
+        next unless res["resourceType"] == "RegulatedAuthorization"
+        Array(res["extension"]).each do |ext|
+          next unless ext["url"].to_s.include?("reimbursementSL")
+          Array(ext["extension"]).each do |sub|
+            sub["valueInteger"] = percent if sub["url"] == "costShare"
+          end
+        end
+      end
+
+      file = Tempfile.new(["cyramza-costshare", ".ndjson"])
+      begin
+        file.write(JSON.generate(bundle))
+        file.flush
+        described_class.new(file.path).to_hash.values.first
+      ensure
+        file.close
+        file.unlink
+      end
+    end
+
+    it "maps the ordinary 10% Selbstbehalt to deductible N" do
+      # The live BAG feed carries costShare 10 for ~97% of all packages, so a
+      # 10 that ends up as "Y" makes every article claim a 40% Selbstbehalt.
+      item = extract_with_cost_share(10)
+      expect(item[:deductible]).to eq("N")
+      expect(item[:deductible20]).to eq("")
+    end
+
+    it "maps the raised 40% Selbstbehalt to deductible Y" do
+      item = extract_with_cost_share(40)
+      expect(item[:deductible]).to eq("Y")
+      expect(item[:deductible20]).to eq("")
+    end
+
+    it "maps the transitional 20% Selbstbehalt to deductible20 Y" do
+      item = extract_with_cost_share(20)
+      expect(item[:deductible20]).to eq("Y")
+    end
+  end
+
   describe Oddb2xml::FhirExtractor, "limitation text resolution" do
     # Build language-variant copies of the Cyramza fixture in-memory:
     # the live FHIR feed never stores limitation text inline, only a
@@ -309,3 +356,32 @@ describe "FHIR Indikationscode support" do
     end
   end
 end
+
+# BAG moved the export on 24.08.2026 and announced it the next evening. The old
+# -latest- alias answers 404 while the old dated snapshots stay in place, so the
+# failure is an import that quietly downloads nothing - the same shape that let
+# ch.oddb.org run for twelve mornings reporting exit 0 and importing nothing.
+describe Oddb2xml::FhirDownloader do
+  it "builds the per-language urls under the publication path" do
+    base = Oddb2xml::FhirDownloader::BASE_URL
+    path = Oddb2xml::FhirDownloader::STATIC_FHIR_PATH
+    prefix = Oddb2xml::FhirDownloader::FILE_PREFIX
+
+    expect(path).to eq "/static/sl/publication/fhir"
+    expect(prefix).to eq "foph-sl-publication"
+    expect(Oddb2xml::FhirDownloader::LANGUAGES).to eq %w[de fr it]
+
+    url = "#{base}#{path}/#{prefix}-latest-de.ndjson"
+    expect(url).to eq "https://epl.bag.admin.ch/static/sl/publication/fhir/" \
+                      "foph-sl-publication-latest-de.ndjson"
+  end
+
+  it "no longer points at the retired export path" do
+    combined = [
+      Oddb2xml::FhirDownloader::STATIC_FHIR_PATH,
+      Oddb2xml::FhirDownloader::FILE_PREFIX
+    ].join(" ")
+    expect(combined).not_to include "foph-sl-export"
+    expect(combined).not_to eq "/static/fhir"
+  end
+end

data/spec/rogger_names_spec.rb

@@ -0,0 +1,88 @@
+# frozen_string_literal: true
+
+require "spec_helper"
+require "oddb2xml/rogger_names"
+
+RSpec.describe Oddb2xml::RoggerNames do
+  let(:rogger_csv) {
+    <<~CSV
+      GTIN,Mediname
+      7680672570037,RINVOQ Ret Tabl 30 mg 28 Stk
+      7680665940021,GABAPENTIN Spirig HC Kaps 100 mg 100 Stk
+      7680625680042,CETIRIZIN Spirig HC Filmtabl 10 mg 10 Stk
+      not-a-gtin,BROKEN ROW
+      7680656721066,
+    CSV
+  }
+
+  describe ".parse" do
+    subject(:map) { described_class.parse(rogger_csv) }
+
+    it "maps 13-digit GTINs to the preferred German name" do
+      expect(map["7680672570037"]).to eq "RINVOQ Ret Tabl 30 mg 28 Stk"
+      expect(map["7680665940021"]).to eq "GABAPENTIN Spirig HC Kaps 100 mg 100 Stk"
+    end
+
+    it "skips rows with a malformed GTIN or an empty name" do
+      expect(map.size).to eq 3
+      expect(map).not_to have_key("7680656721066")
+    end
+
+    it "returns {} for blank input" do
+      expect(described_class.parse("")).to eq({})
+      expect(described_class.parse(nil)).to eq({})
+    end
+  end
+
+  describe ".rogger_csv?" do
+    it "accepts the sheet export header and rejects HTML/empty bodies" do
+      expect(described_class.rogger_csv?(rogger_csv)).to be true
+      expect(described_class.rogger_csv?("\xEF\xBB\xBFGTIN,Mediname\n")).to be true
+      expect(described_class.rogger_csv?(nil)).to be false
+      expect(described_class.rogger_csv?("")).to be false
+      expect(described_class.rogger_csv?("<!DOCTYPE html><html>Sign in - Google Accounts</html>")).to be false
+    end
+  end
+
+  describe ".load" do
+    it "falls back to the bundled data/rogger_liste.csv when the download fails" do
+      flexmock(Oddb2xml::RoggerDownloader).new_instances.should_receive(:download).and_raise(SocketError)
+      map = described_class.load({})
+      expect(map).not_to be_empty
+      expect(map.keys).to all(match(/\A\d{13}\z/))
+      expect(map["7680672570037"]).to eq "RINVOQ Ret Tabl 30 mg 28 Stk"
+    end
+
+    it "falls back to the bundled CSV when the response is not the sheet export" do
+      flexmock(Oddb2xml::RoggerDownloader).new_instances.should_receive(:download)
+        .and_return("<!DOCTYPE html><html>Sign in - Google Accounts</html>")
+      map = described_class.load({})
+      expect(map["7680672570037"]).to eq "RINVOQ Ret Tabl 30 mg 28 Stk"
+    end
+
+    it "never raises, returning {} when no source is available" do
+      flexmock(Oddb2xml::RoggerDownloader).new_instances.should_receive(:download).and_raise(SocketError)
+      flexmock(described_class).should_receive(:source).and_return(nil)
+      expect(described_class.load({})).to eq({})
+    end
+  end
+
+  describe "Builder#apply_rogger_name_overrides!" do
+    it "replaces only the German description of listed GTINs" do
+      builder = Oddb2xml::Builder.new({})
+      builder.rogger_names = {"7680672570037" => "RINVOQ Ret Tabl 30 mg 28 Stk"}
+      builder.refdata = {
+        "7680672570037" => {ean13: "7680672570037", no8: "67257003",
+                            desc_de: "RINVOQ Retardtabletten 30 mg 28 Stk",
+                            desc_fr: "RINVOQ cpr ret 30 mg 28 pce"},
+        "7680671390032" => {ean13: "7680671390032", no8: "67139003",
+                            desc_de: "FAMPYRA Ret Tabl 10 mg 4x14 Stk",
+                            desc_fr: "FAMPYRA cpr ret 10 mg 4x14 pce"}
+      }
+      builder.send(:apply_rogger_name_overrides!)
+      expect(builder.refdata["7680672570037"][:desc_de]).to eq "RINVOQ Ret Tabl 30 mg 28 Stk"
+      expect(builder.refdata["7680672570037"][:desc_fr]).to eq "RINVOQ cpr ret 30 mg 28 pce"
+      expect(builder.refdata["7680671390032"][:desc_de]).to eq "FAMPYRA Ret Tabl 10 mg 4x14 Stk"
+    end
+  end
+end

metadata

@@ -1,14 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: oddb2xml
 version: !ruby/object:Gem::Version
-  version: 3.0.30
+  version: 3.0.32
 platform: ruby
 authors:
 - Yasuhiro Asaka, Zeno R.R. Davatz, Niklaus Giger
-autorequire:
 bindir: bin
 cert_chain: []
-date: 2026-07-11 00:00:00.000000000 Z
+date: 1980-01-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rubyzip
@@ -469,6 +468,7 @@ files:
 - data/gal_groups.yaml
 - data/gtin2ignore.yaml
 - data/product_overrides.yaml
+- data/rogger_liste.csv
 - data/wala_arzneimittel.csv
 - data/weleda_arzneimittel.csv
 - dokumentation_calc.textile
@@ -490,6 +490,7 @@ files:
 - lib/oddb2xml/parslet_compositions.rb
 - lib/oddb2xml/proxy_check.rb
 - lib/oddb2xml/refdata_cleanup.rb
+- lib/oddb2xml/rogger_names.rb
 - lib/oddb2xml/semantic_check.rb
 - lib/oddb2xml/util.rb
 - lib/oddb2xml/version.rb
@@ -499,8 +500,12 @@ files:
 - oddb2xml.xsd
 - oddb_calc.xsd
 - scripts/generate_index_html.sh
+- scripts/get_transfer.sh
 - scripts/run_oddb2xml.sh
+- scripts/setup_aips2sqlite.sh
 - scripts/setup_mediupdatexml_web.sh
+- scripts/setup_new_server.sh
+- scripts/setup_rust2xml.sh
 - scripts/swissmedic_watch.sh
 - scripts/transfer.sh
 - scripts/visitor_stats.py
@@ -580,6 +585,7 @@ files:
 - spec/parslet_spec.rb
 - spec/proxy_check_spec.rb
 - spec/refdata_cleanup_spec.rb
+- spec/rogger_names_spec.rb
 - spec/spec_helper.rb
 - spec/weleda_sl_spec.rb
 - test_options.rb
@@ -591,7 +597,6 @@ homepage: https://github.com/zdavatz/oddb2xml
 licenses:
 - GPL-3.0-only
 metadata: {}
-post_install_message:
 rdoc_options: []
 require_paths:
 - lib
@@ -606,8 +611,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.5.22
-signing_key:
+rubygems_version: 3.6.7
 specification_version: 4
 summary: oddb2xml creates xml files.
 test_files:
@@ -687,5 +691,6 @@ test_files:
 - spec/parslet_spec.rb
 - spec/proxy_check_spec.rb
 - spec/refdata_cleanup_spec.rb
+- spec/rogger_names_spec.rb
 - spec/spec_helper.rb
 - spec/weleda_sl_spec.rb