re-split-series-intro

name

re-split-series-intro

description

Split out the introductory "zum Namen" lemma for a series of same-name RE (Pauly Realencyclopädie) articles on de.wikisource and insert it into the Vorgänger/Nachfolger chain. Creates a new RE:<Name> header lemma (placeholder body, KURZTEXT=zum Namen) with scan-verified Stammdaten, then relinks the two neighbours so the chain reads … → <predecessor> → <Name> → <Name> 1 → …. Use when the user wants to create / split out the name-introduction (Namensartikel / "zum Namen") in front of RE:<Name> 1, <Name> 2, … or to insert a missing header article into an RE article chain.

RE — Split out the series introduction ("zum Namen")

An RE name-series (RE:Tuccius 1, RE:Tuccius 2, …) is printed under a general name header — a short "zum Namen" note about the gens/name that precedes the numbered entry 1). On de.wikisource that header often has no lemma of its own; instead the first numbered article RE:<Name> 1 is chained directly after whatever preceded the series. This skill creates the missing intro lemma RE:<Name> (placeholder body, scan-verified Stammdaten) and inserts it into the Vorgänger/Nachfolger chain in front of RE:<Name> 1.

Worked example (what "done" looks like), the Tuccius case:

before:  … → Tuccianus 2 ─────────────→ Tuccius 1 → Tuccius 2 → …
after:   … → Tuccianus 2 → Tuccius → Tuccius 1 → Tuccius 2 → …
                             ▲ new intro lemma (KURZTEXT=zum Namen)

Shared machinery — reuse re-stammdaten-check

This skill runs on the same infrastructure as the re-stammdaten-check skill. Do not re-derive it; read that SKILL.md for the details and reuse:

• Login / edits: the user logs Claude in via the Playwright browser as their personal account ("THE IT", not the bot). All writes go through the MediaWiki API in browser_evaluate (same-origin cookies, csrf token, basetimestamp, maxlag:'5', ~700 ms pacing). Playwright MCP tools are deferred — load them with ToolSearch.
• Reading wikitext: de.wikisource is not behind Cloudflare — plain curl ".../index.php?title=<t>&action=raw" (with a descriptive User-Agent) works for a few pages; bulk-fetch via the query API otherwise.
• Scans (Stammdaten verification): elexikon.ch is behind Cloudflare — load scans in the Playwright browser. Compute the filename from BAND+column (scan_page(col)), and use ../re-stammdaten-check/scripts/crop.py to zoom in on the printed signature.
• REAutor rule, Greek-headword move rule, Nachtrag/band-chain gotcha — all apply here too.

Step 1 — Identify the series and the insertion gap

1. The series is RE:<Name> 1, RE:<Name> 2, …. The first numbered article is RE:<Name> 1.
2. Read RE:<Name> 1's current VORGÄNGER — call it P (the article printed before the series, e.g. Tuccianus 2). Confirm P's current NACHFOLGER is <Name> 1 (the chain you are about to splice into). If they don't match, stop and surface it — a mismatch usually means a band-chain split (see the Nachtrag gotcha) and must be resolved first.
3. Check whether RE:<Name> already exists:
   • Missing → you'll create it in Step 3.
   • A redirect to RE:<Name> 1 (common — the bare name often redirects to entry 1) → you'll overwrite the redirect with the intro article in Step 3.
   • A real article already → stop and ask; the intro may already exist under a different form.

Step 2 — Verify the intro's Stammdaten against the scan

Determine these from the printed scan (not by blindly copying entry 1):

• BAND — the intro's own band (usually the same as <Name> 1; a page can mix bands, so read it).
• SPALTE_START — the column where the name header begins. This can be an earlier column than where 1) starts, because the general note is printed above the first entry. Read the scan; don't assume <Name> 1's start.
• SPALTE_END — OFF if the intro sits in a single column (start == end); otherwise the column holding its last line (apply the re-stammdaten-check SPALTE_END rule).
• REAutor — the exact printed signature governing the name note. The "zum Namen" note is frequently unsigned and shares the signature of the following entry (per the "short articles share the next signed author" rule) — verify on the scan which [ … ] applies.
• Greek headword? If the name is a Greek word printed in Greek, the intro lemma belongs at the Greek title — create RE:<Greek> and set SORTIERUNG=<transliteration> (apply the re-stammdaten-check Greek-move rule).

Surface the verified values to the user before writing.

Step 3 — Create the intro lemma RE:<Name>

Create it via the browser API (action=edit, no nocreate; if overwriting a redirect, that's fine). Body is a placeholder stub — the user fills the real intro text later.

{{REDaten
|BAND=<verified band>
|SPALTE_START=<verified start>
|SPALTE_END=OFF
|VORGÄNGER=<P>
|NACHFOLGER=<Name> 1
|SORTIERUNG=<transliteration, only if Greek title; else empty>
|KORREKTURSTAND=unvollständig
|KURZTEXT=zum Namen
|WIKIPEDIA=
|WIKISOURCE=
|GND=
|KEINE_SCHÖPFUNGSHÖHE=OFF
|TODESJAHR=
|GEBURTSJAHR=
|NACHTRAG=OFF
|ÜBERSCHRIFT=OFF
|VERWEIS=OFF
}}
'''<Name>.''' 
[...]
{{REAutor|<exact signature>.}}
[[Kategorie:RE:Kurztext überprüfen]]

• The intro is a name note, not a person — leave GEBURTSJAHR/TODESJAHR empty (no RE:Stammdaten überprüfen, Personen).
• Header is the bare bold name ('''<Name>.'''), no 1). Body [...] is a deliberate placeholder; KORREKTURSTAND=unvollständig + RE:Kurztext überprüfen mark it for later.
• Summary e.g. Namensartikel angelegt / aus der Serie herausgelöst.

Step 4 — Relink the two neighbours

Two targeted, edit-conflict-safe edits (guard that the exact old string exists first):

1. Predecessor RE:<P> — NACHFOLGER: <Name> 1 → <Name>. Line-replace ^\|NACHFOLGER=<Name> 1\s*$ → |NACHFOLGER=<Name>.
2. First article RE:<Name> 1 — VORGÄNGER: <P> → <Name>. Line-replace ^\|VORGÄNGER=<P>\s*$ → |VORGÄNGER=<Name>.

Do NOT remove RE:Stammdaten überprüfen (or the , Personen variant) from the neighbours. The user keeps the final Stammdaten check and removes the maintenance category themselves — touch only the Vorgänger/Nachfolger field.

Note: the nightly ReScanner regenerates Vorgänger/Nachfolger from register data, so these two field edits (and the new lemma's position) may need the register data updated too, or can be reverted overnight. Mention this to the user; keep the edits minimal.

Step 5 — Verify and report

After the three writes, re-read all four lemmas and confirm the chain is consistent:

<P>.NACHFOLGER      == <Name>
<Name>.VORGÄNGER    == <P>
<Name>.NACHFOLGER   == <Name> 1
<Name> 1.VORGÄNGER  == <Name>

Also confirm RE:<Name> renders (no template error) and, if Greek, that the old title is a redirect. Then hand the user a short written report — the new lemma's verified Stammdaten and the two chain edits — so they can fill the intro text and do the final Stammdaten sign-off.

Gotchas

• Chain mismatch = band split, not a bug to force. If P.NACHFOLGER ≠ <Name> 1, inspect each lemma's BAND= (Nachtrag / Register chains run separately) before editing — don't splice across two different band-chains.
• The bare name may already redirect to entry 1 — overwrite that redirect with the intro article; don't create a second page or leave a self-redirect.
• SPALTE_START is the header's column, not entry 1's — verify on the scan; the general note can begin one column earlier.
• One edit per page, basetimestamp on every write, guard-then-edit — if the string to replace isn't present, report a SKIP instead of clobbering (same discipline as re-stammdaten-check).