cocina-models 0.119.0 → 0.121.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 95fdb985d82245efee3898558df726439ee7d27e7fa45bd47815b44d52f58679
-  data.tar.gz: 0a7644f423957828280316d0c5709847f260919e1ceb384c130e8196f4d5eb6a
+  metadata.gz: cfcb2c2a393845b1c9129ab48f5fc7340beb8067936694798e800a5935bf1ac9
+  data.tar.gz: f9dc80c520ddecb670a5df20273c635146f75d75025132402d4d1329090e29fd
 SHA512:
-  metadata.gz: ce4a6079617e05fd8d5b1547d16e983e06f0f53914339f9f19dcb4a870d52c1c565844bb39b0ed5decffe2c68f2a4fc4f2b385b9d8dd4796700f2c1f9aed5c7b
-  data.tar.gz: eb72b867dcab3ad0ee2e63d3b68276c17fb831365f4c6a785a95fa8250998b663c6bcca2d50d4ffc6786044a17df7321847c069595587a9ae3ac96120f093020
+  metadata.gz: 2f7ccadc053ba81117275251401432ec92c65fefd86da94c25e7c862f8680c868d3f392a84b3966fb48609a5c383abd8dc90b8a51e457fd2fe8d3792371bd4a6
+  data.tar.gz: c1355245a0571f8933155c9de04b38226b55b96a3ffdb899af3d2af553e30f53ae7986a41e6293339a90cc2a017fcd6eec273fcd1ce64df57b26cbbb5f161193

data/.circleci/config.yml

@@ -1,6 +1,6 @@
 version: 2.1
 orbs:
-  ruby-rails: sul-dlss/ruby-rails@4.8.0
+  ruby-rails: sul-dlss/ruby-rails@4.10.0
 workflows:
   build:
     jobs:

data/.claude/skills/cocina-jq-query/SKILL.md

@@ -0,0 +1,8 @@
+---
+name: cocina-jq-query
+description: Build and validate a jq query against a Cocina Model JSON serialization. Use when the user wants to query, filter, or transform a Cocina object (DRO, Collection, AdminPolicy) using jq, or asks for help writing a jq expression for Cocina data.
+---
+
+# cocina-jq-query
+
+Follow the workflow defined in [AGENTS.md](../../../AGENTS.md) under **cocina-jq**.

data/.gitignore

@@ -9,3 +9,8 @@
 
 # rspec failure tracking
 .rspec_status
+
+*.jq.txt
+*.jsonl.xz
+*.csv
+*-playground.html

data/.rubocop.yml

@@ -126,6 +126,12 @@ RSpec/BeEq: # new in 2.9.0
 RSpec/BeNil: # new in 2.9.0
   Enabled: true
 
+RSpec/DiscardedMatcher: # new in 3.10
+  Enabled: true
+
+RSpec/MatchWithSimpleRegex: # new in 3.10
+  Enabled: true
+
 RSpec/MultipleExpectations:
   Enabled: false
 

data/AGENTS.md

@@ -0,0 +1,208 @@
+# Agent Instructions for cocina-models
+
+## cocina-jq — Build jq queries for Cocina Model JSON
+
+Use this workflow when the user wants to query, filter, or transform a Cocina object (DRO, Collection, AdminPolicy) using jq.
+
+### Step 1 — Check prerequisites
+
+#### Check jq
+Run `jq --version`. If jq is not installed, tell the user:
+> `jq` is not installed. Install it with `brew install jq`, then retry.
+
+Stop here if jq is missing.
+
+#### Check parallel
+Run `parallel --version`. If parallel is not installed, tell the user:
+> `parallel` is not installed. Install it with `brew install parallel`, then retry.
+
+#### Check pv
+Run `pv --version`. If pv is not installed, tell the user:
+> `pv` is not installed. Install it with `brew install pv`, then retry.
+
+### Output format (always apply)
+
+Every jq query produced by this skill **must output a CSV line** using `@csv`. The **first field must always be the external identifier** (`externalIdentifier`). Additional fields follow based on the user's query. Example:
+
+```
+"druid:bc123df4567","some value","another value"
+```
+
+Use `[.externalIdentifier, ...] | @csv` as the output expression. Apply this constraint automatically — do not ask the user whether to include the external identifier.
+
+### Step 2 — Resume or collect inputs
+
+First, ask the user:
+
+> Do you want to resume an existing query?
+
+**If yes:** Ask for the filename of the `.jq.txt` file (e.g., `contributor-name-uri-non-loc.jq.txt`). Read that file from the project root. The file header contains the original inputs as comments (query description, expected output description, example JSON, example output). Parse those comments to reconstruct the inputs. Confirm with the user that the loaded values look correct, then proceed to Step 3 with those inputs (skip re-asking for them).
+
+**If no:** Ask the user for each input, one at a time:
+
+1. **Query description** — what should the query do? (e.g., "extract all file labels from structural")
+2. **Expected output description** — what additional values (beyond the external identifier) should appear in the output?
+3. **Example Cocina object** — paste JSON directly
+4. **Example output** — paste the exact expected CSV output (must start with the external identifier as the first field)
+
+Explicitly ask for each input; do not infer or guess.
+
+### Step 3 — Clarify ambiguities
+
+Review the inputs from Step 2. If anything is unclear or underspecified, ask the user targeted questions before proceeding. Examples of things to clarify:
+
+- Is the query meant to return one value per object, or aggregate across many objects?
+- Should missing or null fields be skipped, returned as null, or cause an error?
+- Are there edge cases in the data structure the query must handle (e.g., empty arrays, nested arrays, optional fields)?
+- Is the output format exactly as shown, or is there flexibility (e.g., flat vs. nested)?
+
+Ask only questions that would change how the query is written. Do not ask about things already clear from the inputs. If everything is unambiguous, skip this step silently and proceed.
+
+### Step 4 — Load relevant schema portion
+
+Read `schema.json` from the project root. Extract only the `$defs` entries relevant to the Cocina object type found in the example's `type` field:
+
+- `https://cocina.sul.stanford.edu/models/object` → DRO-related defs
+- `https://cocina.sul.stanford.edu/models/collection` → Collection-related defs
+- `https://cocina.sul.stanford.edu/models/admin_policy` → AdminPolicy-related defs
+
+Include only the defs actually referenced (follow `$ref` chains up to 2 levels deep). Do not load the entire schema.
+
+### Step 5 — Generate and validate the query (up to 3 attempts)
+
+**Attempt 1:** Use the schema excerpt, example JSON, query description, and expected output to write a jq query.
+
+Run it:
+```bash
+echo '<example_json>' | jq '<query>'
+```
+
+Compare actual output to the example output. If it matches → proceed to Step 5.
+
+**Attempt 2 (if attempt 1 fails):** Run `man jq` to load the jq manual. Use it to refine the query. Re-run and validate.
+
+**Attempt 3 (if attempt 2 fails):** Make a final attempt using all context. Re-run and validate.
+
+**After 3 failures:** Present the best attempt, explain what is wrong, and ask the user to clarify.
+
+### Step 6 — Generate local HTML playground
+
+Write `<slug>-playground.html` in the project root using the template below.
+
+Substitute:
+- Every occurrence of `SLUG` → the actual slug string
+- `JSON_PLACEHOLDER` → the example JSON (pretty-printed) passed through `JSON.stringify` a second time, producing a valid JS string literal (e.g. `"{\"foo\":\"bar\"}"`)
+- `QUERY_PLACEHOLDER` → the validated jq query passed through `JSON.stringify`, producing a valid JS string literal (e.g. `".foo"`)
+
+```html
+<!DOCTYPE html>
+<html lang="en">
+<head>
+  <meta charset="UTF-8">
+  <title>jq playground — SLUG</title>
+  <style>
+    body { font-family: monospace; margin: 2rem; background: #1e1e1e; color: #d4d4d4; }
+    h2 { color: #9cdcfe; }
+    textarea, input { width: 100%; box-sizing: border-box; background: #252526; color: #d4d4d4; border: 1px solid #444; padding: 8px; font-family: monospace; font-size: 13px; border-radius: 3px; }
+    textarea { height: 260px; resize: vertical; }
+    input { height: 36px; }
+    button { margin-top: 8px; background: #0e639c; color: white; border: none; padding: 8px 20px; cursor: pointer; font-size: 14px; border-radius: 3px; }
+    button:hover { background: #1177bb; }
+    label { display: block; margin-top: 16px; margin-bottom: 4px; font-size: 12px; color: #9cdcfe; text-transform: uppercase; letter-spacing: 0.05em; }
+    #output { background: #252526; border: 1px solid #444; padding: 12px; min-height: 80px; white-space: pre-wrap; word-break: break-all; border-radius: 3px; }
+    .error { color: #f44747; }
+  </style>
+</head>
+<body>
+  <h2>jq playground — SLUG</h2>
+  <label>JSON Input</label>
+  <textarea id="json"></textarea>
+  <label>jq Filter</label>
+  <input id="query" type="text" />
+  <button onclick="run()">&#9654; Run</button>
+  <label>Output</label>
+  <pre id="output">(click Run)</pre>
+
+  <script src="https://cdn.jsdelivr.net/npm/jq-web@0.5.1/jq.wasm.js"></script>
+  <script>
+    const INITIAL_JSON = JSON_PLACEHOLDER;
+    const INITIAL_QUERY = QUERY_PLACEHOLDER;
+    document.getElementById('json').value = JSON.stringify(JSON.parse(INITIAL_JSON), null, 2);
+    document.getElementById('query').value = INITIAL_QUERY;
+
+    function run() {
+      const json = document.getElementById('json').value;
+      const query = document.getElementById('query').value;
+      const out = document.getElementById('output');
+      out.className = '';
+      out.textContent = 'Running…';
+      jq.promised.raw(json, query)
+        .then(r => { out.textContent = r || '(empty output)'; })
+        .catch(e => { out.className = 'error'; out.textContent = String(e); });
+    }
+  </script>
+</body>
+</html>
+```
+
+### Step 7 — Write the query to a .jq.txt file
+
+Generate a short kebab-case slug summarizing the query (e.g., `invalid-encoding`, `file-label-extract`). Write the file `<slug>.jq.txt` in the project root with the following structure:
+
+1. A comment header containing the user's inputs:
+
+```
+# Query description: <query description from Step 2>
+#
+# Expected output: <expected output description from Step 2>
+#
+# Example input:
+# <example Cocina JSON, each line prefixed with "# ">
+#
+# Example output:
+# <example output CSV, each line prefixed with "# ">
+```
+
+2. A blank line, then the validated jq query.
+
+The comment lines must use `#` so the file remains valid jq syntax. When resuming (Step 2 resume path), parse these comment sections by their labels to reconstruct the inputs.
+
+Find the most recent `.jsonl.xz` file in the project root by listing `*.jsonl.xz` files sorted by name descending and taking the first result.
+
+### Step 8 — Output
+
+Present:
+1. The jq query in a code block
+2. A 1–3 sentence explanation of how it works
+3. A markdown link to the local playground file using a `file://` URL (e.g. `[Open playground](file:///Users/someuser/data/sdr/cocina-models/<slug>-playground.html)`) — substitute the actual absolute path — plus the equivalent shell command (`open <slug>-playground.html`) for reference
+4. A ready-to-run shell snippet:
+
+```
+xzcat <most-recent .jsonl.xz filename> \
+  | pv -l -s 5500000 \
+  | parallel -j$(sysctl -n hw.logicalcpu) --pipe --block 50M --recend '\n' \
+              jq -rcf <slug>.jq.txt \
+  | bundle exec bin/enhance-report-csv \
+  | tee <slug>.csv
+```
+
+Substitute the actual filenames — do not leave placeholders.
+
+Also, remind the user to tunnel to Solr in a separate terminal with:
+```
+ssh -L 8990:sul-solr-prod-a.stanford.edu:80 lyberadmin@argo-prod-02.stanford.edu
+```
+
+### Step 9 — Iterate
+
+After presenting step 7 output, prompt the user:
+
+> Want to refine the query? You can describe a change (e.g., "also filter by `type`") or paste a modified jq expression directly.
+
+**If the user describes a change:** Update the query to satisfy the new requirement, re-run against the example JSON (same validation loop as Step 5, up to 3 attempts), then repeat Steps 6–8 with the updated query and slug.
+
+**If the user pastes a modified query directly:** Validate it by running against the example JSON. If it produces valid output, skip straight to repeating Steps 6–8. If it errors, diagnose and fix (up to 3 attempts), then repeat Steps 5–7.
+
+**Each iteration overwrites the `.jq.txt` file and `<slug>-playground.html`** (same slug unless the query purpose changed significantly, in which case generate a new slug) **and replaces all previous outputs** with updated versions.
+
+Continue offering to iterate after each round until the user is satisfied.

data/Gemfile

@@ -5,6 +5,8 @@ source 'https://rubygems.org'
 # Specify your gem's dependencies in cocina-models.gemspec
 gemspec
 
+gem 'csv'
 gem 'debug'
+gem 'rsolr'
 gem 'rspec_junit_formatter' # For CircleCI
 gem 'ruby-progressbar'

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    cocina-models (0.119.0)
+    cocina-models (0.121.0)
       activesupport
       deprecation
       dry-struct (~> 1.0)
@@ -9,8 +9,7 @@ PATH
       edtf
       equivalent-xml
       i18n
-      json_schemer (~> 2.0)
-      jsonpath
+      jsonschema_rs
       nokogiri
       super_diff
       thor
@@ -36,8 +35,10 @@ GEM
     attr_extras (7.1.0)
     base64 (0.3.0)
     bigdecimal (4.1.2)
+    builder (3.3.0)
     concurrent-ruby (1.3.6)
     connection_pool (3.0.2)
+    csv (3.3.5)
     date (3.5.1)
     debug (1.11.1)
       irb (~> 1.10)
@@ -74,7 +75,12 @@ GEM
     equivalent-xml (0.6.0)
       nokogiri (>= 1.4.3)
     erb (6.0.4)
-    hana (1.3.7)
+    faraday (2.14.2)
+      faraday-net_http (>= 2.0, < 3.5)
+      json
+      logger
+    faraday-net_http (3.4.4)
+      net-http (~> 0.5)
     i18n (1.14.8)
       concurrent-ruby (~> 1.0)
     ice_nine (0.11.2)
@@ -84,21 +90,19 @@ GEM
       prism (>= 1.3.0)
       rdoc (>= 4.0.0)
       reline (>= 0.4.2)
-    json (2.19.7)
-    json_schemer (2.5.0)
-      bigdecimal
-      hana (~> 1.3)
-      regexp_parser (~> 2.0)
-      simpleidn (~> 0.2)
-    jsonpath (1.1.5)
-      multi_json
+    json (2.19.8)
+    jsonschema_rs (0.46.5-arm64-darwin)
+      bigdecimal (>= 3.1, < 5)
+    jsonschema_rs (0.46.5-x86_64-linux)
+      bigdecimal (>= 3.1, < 5)
     language_server-protocol (3.17.0.5)
     lint_roller (1.1.0)
     logger (1.7.0)
     minitest (6.0.6)
       drb (~> 2.0)
       prism (~> 1.5)
-    multi_json (1.21.1)
+    net-http (0.9.1)
+      uri (>= 0.11.1)
     nokogiri (1.19.3-arm64-darwin)
       racc (~> 1.4)
     nokogiri (1.19.3-x86_64-linux-gnu)
@@ -114,7 +118,7 @@ GEM
       prettyprint
     prettyprint (0.2.0)
     prism (1.9.0)
-    psych (5.3.1)
+    psych (5.4.0)
       date
       stringio
     racc (1.8.1)
@@ -127,6 +131,9 @@ GEM
     regexp_parser (2.12.0)
     reline (0.6.3)
       io-console (~> 0.5)
+    rsolr (2.6.0)
+      builder (>= 2.1.2)
+      faraday (>= 0.9, < 3, != 2.0.0)
     rspec (3.13.2)
       rspec-core (~> 3.13.0)
       rspec-expectations (~> 3.13.0)
@@ -159,9 +166,10 @@ GEM
     rubocop-rake (0.7.1)
       lint_roller (~> 1.1)
       rubocop (>= 1.72.1)
-    rubocop-rspec (3.9.0)
+    rubocop-rspec (3.10.2)
       lint_roller (~> 1.1)
-      rubocop (~> 1.81)
+      regexp_parser (>= 2.0)
+      rubocop (~> 1.86, >= 1.86.2)
     ruby-progressbar (1.13.0)
     securerandom (0.4.1)
     simplecov (0.22.0)
@@ -170,7 +178,6 @@ GEM
       simplecov_json_formatter (~> 0.1)
     simplecov-html (0.13.2)
     simplecov_json_formatter (0.1.4)
-    simpleidn (0.2.3)
     stringio (3.2.0)
     super_diff (0.19.0)
       attr_extras (>= 6.2.4, < 8)
@@ -193,8 +200,10 @@ PLATFORMS
 DEPENDENCIES
   bundler (>= 2.0, < 5)
   cocina-models!
+  csv
   debug
   rake (~> 13.0)
+  rsolr
   rspec (~> 3.0)
   rspec_junit_formatter
   rubocop (~> 1.24)
@@ -209,10 +218,12 @@ CHECKSUMS
   attr_extras (7.1.0) sha256=d96fc9a9dd5d85ba2d37762440a816f840093959ae26bb90da994c2d9f1fc827
   base64 (0.3.0) sha256=27337aeabad6ffae05c265c450490628ef3ebd4b67be58257393227588f5a97b
   bigdecimal (4.1.2) sha256=53d217666027eab4280346fba98e7d5b66baaae1b9c3c1c0ffe89d48188a3fbd
-  bundler (4.0.12) sha256=7f8b757d28dfb636e7b24fba2344ac6dd13b5b24f4b46d62573d483f211825ac
-  cocina-models (0.119.0)
+  builder (3.3.0) sha256=497918d2f9dca528fdca4b88d84e4ef4387256d984b8154e9d5d3fe5a9c8835f
+  bundler (4.0.13) sha256=19f08be7f27022cf0b89f27da0b044ae075e8270a9ef44ad248a932614e1ca3b
+  cocina-models (0.121.0)
   concurrent-ruby (1.3.6) sha256=6b56837e1e7e5292f9864f34b69c5a2cbc75c0cf5338f1ce9903d10fa762d5ab
   connection_pool (3.0.2) sha256=33fff5ba71a12d2aa26cb72b1db8bba2a1a01823559fb01d29eb74c286e62e0a
+  csv (3.3.5) sha256=6e5134ac3383ef728b7f02725d9872934f523cb40b961479f69cf3afa6c8e73f
   date (3.5.1) sha256=750d06384d7b9c15d562c76291407d89e368dda4d4fff957eb94962d325a0dc0
   debug (1.11.1) sha256=2e0b0ac6119f2207a6f8ac7d4a73ca8eb4e440f64da0a3136c30343146e952b6
   deprecation (1.1.0) sha256=01707cea9a6ed2d7270377457941f43394a345e6dd8048e1be6d18ff2f2a01e1
@@ -227,19 +238,20 @@ CHECKSUMS
   edtf (3.2.0) sha256=a15a0ee274e49c8047a3ebb5d61d793ba44f7f8ffbf0595392c467e3ea8d2447
   equivalent-xml (0.6.0) sha256=8919761efa848ad0846369ff8be1f646b17e5061698c4867b09829000cc3f487
   erb (6.0.4) sha256=38e3803694be357fe2bfe312487c74beaf9fb4e5beb3e22498952fe1645b95d9
-  hana (1.3.7) sha256=5425db42d651fea08859811c29d20446f16af196308162894db208cac5ce9b0d
+  faraday (2.14.2) sha256=73ccb9994a9e8648f010e32eca2ae82e41c57860aa10932cda29418b9e0223ad
+  faraday-net_http (3.4.4) sha256=0e78af151747ed1b00f33e25973b4bc220d7f16c00c39676817c8b12331eb588
   i18n (1.14.8) sha256=285778639134865c5e0f6269e0b818256017e8cde89993fdfcbfb64d088824a5
   ice_nine (0.11.2) sha256=5d506a7d2723d5592dc121b9928e4931742730131f22a1a37649df1c1e2e63db
   io-console (0.8.2) sha256=d6e3ae7a7cc7574f4b8893b4fca2162e57a825b223a177b7afa236c5ef9814cc
   irb (1.18.0) sha256=de9454a0703a54704b9811a5ef31a60c86949fbf4013fcf244fabc7c775248e3
-  json (2.19.7) sha256=fe432c8639f6efff69f9d73b518a3705d9581ab93156f981ea72806e1e5bcc3e
-  json_schemer (2.5.0) sha256=2f01fb4cce721a4e08dd068fc2030cffd0702a7f333f1ea2be6e8991f00ae396
-  jsonpath (1.1.5) sha256=29f70467193a2dc93ab864ec3d3326d54267961acc623f487340eb9c34931dbe
+  json (2.19.8) sha256=6354310fd76ef69b87d5bd1f38b40d730613baf90b6803d2d0a48f618d32dfaa
+  jsonschema_rs (0.46.5-arm64-darwin) sha256=e80414ed67f0956d3e06474a2fa076fc4a7b722f00e5d7142b70289c016ac6f1
+  jsonschema_rs (0.46.5-x86_64-linux) sha256=345c65ec7a5abf8879b9c9356752f0fdf4c9926f6480458fc32803a871b5cbb3
   language_server-protocol (3.17.0.5) sha256=fd1e39a51a28bf3eec959379985a72e296e9f9acfce46f6a79d31ca8760803cc
   lint_roller (1.1.0) sha256=2c0c845b632a7d172cb849cc90c1bce937a28c5c8ccccb50dfd46a485003cc87
   logger (1.7.0) sha256=196edec7cc44b66cfb40f9755ce11b392f21f7967696af15d274dde7edff0203
   minitest (6.0.6) sha256=153ea36d1d987a62942382b61075745042a2b3123b1cd48f4c3675af9cc7d6f1
-  multi_json (1.21.1) sha256=e6126a31808e3b4d19f483c775ceac34df190dffa62adfb63a165ee14ba68080
+  net-http (0.9.1) sha256=25ba0b67c63e89df626ed8fac771d0ad24ad151a858af2cc8e6a716ca4336996
   nokogiri (1.19.3-arm64-darwin) sha256=71b9bd424b1b7abc18b05052a1a3cfd3627abdca62be280854cc411791357e42
   nokogiri (1.19.3-x86_64-linux-gnu) sha256=2f5078620fe12e83669b5b17311b32532a8153d02eee7ad06948b926d6080976
   optimist (3.2.1) sha256=8cf8a0fd69f3aa24ab48885d3a666717c27bc3d9edd6e976e18b9d771e72e34e
@@ -249,13 +261,14 @@ CHECKSUMS
   pp (0.6.3) sha256=2951d514450b93ccfeb1df7d021cae0da16e0a7f95ee1e2273719669d0ab9df6
   prettyprint (0.2.0) sha256=2bc9e15581a94742064a3cc8b0fb9d45aae3d03a1baa6ef80922627a0766f193
   prism (1.9.0) sha256=7b530c6a9f92c24300014919c9dcbc055bf4cdf51ec30aed099b06cd6674ef85
-  psych (5.3.1) sha256=eb7a57cef10c9d70173ff74e739d843ac3b2c019a003de48447b2963d81b1974
+  psych (5.4.0) sha256=14f72d69a611af663d7d70e4a7b67d9eb1f3ae9f8d916b478961d5a0075ba5b7
   racc (1.8.1) sha256=4a7f6929691dbec8b5209a0b373bc2614882b55fc5d2e447a21aaa691303d62f
   rainbow (3.1.1) sha256=039491aa3a89f42efa1d6dec2fc4e62ede96eb6acd95e52f1ad581182b79bc6a
   rake (13.4.2) sha256=cb825b2bd5f1f8e91ca37bddb4b9aaf345551b4731da62949be002fa89283701
   rdoc (7.2.0) sha256=8650f76cd4009c3b54955eb5d7e3a075c60a57276766ebf36f9085e8c9f23192
   regexp_parser (2.12.0) sha256=35a916a1d63190ab5c9009457136ae5f3c0c7512d60291d0d1378ba18ce08ebb
   reline (0.6.3) sha256=1198b04973565b36ec0f11542ab3f5cfeeec34823f4e54cebde90968092b1835
+  rsolr (2.6.0) sha256=4b3bcea772cac300562775c20eeddedf63a6b7516a070cb6fbde000b09cfe12b
   rspec (3.13.2) sha256=206284a08ad798e61f86d7ca3e376718d52c0bc944626b2349266f239f820587
   rspec-core (3.13.6) sha256=a8823c6411667b60a8bca135364351dda34cd55e44ff94c4be4633b37d828b2d
   rspec-expectations (3.13.5) sha256=33a4d3a1d95060aea4c94e9f237030a8f9eae5615e9bd85718fe3a09e4b58836
@@ -265,13 +278,12 @@ CHECKSUMS
   rubocop (1.87.0) sha256=b9d9ddf55116a513f8ef2c7ae660662d8b49301f118d3f0df61865b33a5c188d
   rubocop-ast (1.49.1) sha256=4412f3ee70f6fe4546cc489548e0f6fcf76cafcfa80fa03af67098ffed755035
   rubocop-rake (0.7.1) sha256=3797f2b6810c3e9df7376c26d5f44f3475eda59eb1adc38e6f62ecf027cbae4d
-  rubocop-rspec (3.9.0) sha256=8fa70a3619408237d789aeecfb9beef40576acc855173e60939d63332fdb55e2
+  rubocop-rspec (3.10.2) sha256=0b3e2ecc592cd10ecbf0095bb58d1e357905276e069643523cc19eb7495f65e2
   ruby-progressbar (1.13.0) sha256=80fc9c47a9b640d6834e0dc7b3c94c9df37f08cb072b7761e4a71e22cff29b33
   securerandom (0.4.1) sha256=cc5193d414a4341b6e225f0cb4446aceca8e50d5e1888743fac16987638ea0b1
   simplecov (0.22.0) sha256=fe2622c7834ff23b98066bb0a854284b2729a569ac659f82621fc22ef36213a5
   simplecov-html (0.13.2) sha256=bd0b8e54e7c2d7685927e8d6286466359b6f16b18cb0df47b508e8d73c777246
   simplecov_json_formatter (0.1.4) sha256=529418fbe8de1713ac2b2d612aa3daa56d316975d307244399fa4838c601b428
-  simpleidn (0.2.3) sha256=08ce96f03fa1605286be22651ba0fc9c0b2d6272c9b27a260bc88be05b0d2c29
   stringio (3.2.0) sha256=c37cb2e58b4ffbd33fe5cd948c05934af997b36e0b6ca6fdf43afa234cf222e1
   super_diff (0.19.0) sha256=c35fc1c0daa223d67b203fe3fb49a6cfd67850a53920319565c3c654e03ec902
   thor (1.5.0) sha256=e3a9e55fe857e44859ce104a84675ab6e8cd59c650a49106a05f55f136425e73
@@ -283,4 +295,4 @@ CHECKSUMS
   zeitwerk (2.8.2) sha256=7212a61311083c604184b1ea2574b9aa05cd14f855a0841c06985cabe9181d12
 
 BUNDLED WITH
-  4.0.12
+  4.0.13

data/README.md

@@ -51,6 +51,9 @@ exe/generator generate_vocab
 exe/generator generate_descriptive_docs
 ```
 
+## Reports / querying
+jq-based queries can be authored against a [local data export](https://github.com/sul-dlss/dor-services-app#export-data) using the `/cocina-jq-query` skill. This will help with constructing and efficiently running the query.
+
 ## Testing
 
 The generator is tested via its output when run against `schema.json`, viz., the Cocina model classes. Thus, `generate` should be run after any changes to `schema.json`.
@@ -158,7 +161,7 @@ This list of services is known to include:
 * [sul-dlss/sdr-api](https://github.com/sul-dlss/sdr-api)
 * [sul-dlss/dor-services-app](https://github.com/sul-dlss/dor-services-app/)
 
-Perform `bundle update --conservative cocina-models dor-services-client` in the services above and make PRs for those repos. You may first need to update how these gems are pinned in the `Gemfile` in order to bump them.
+Perform `bundle update cocina-models dor-services-client --conservative` in the services above and make PRs for those repos. You may first need to update how these gems are pinned in the `Gemfile` in order to bump them.
 
 Get the directly coupled services PRs merged before the deploy in step 5.
 

data/bin/enhance-report-csv

@@ -0,0 +1,90 @@
+#!/usr/bin/env ruby
+# frozen_string_literal: true
+
+require 'bundler/setup'
+require 'csv'
+require 'optparse'
+require 'rsolr'
+
+# This script reads a CSV from standard input, where the first column is expected to be a druid.
+# It queries Solr for each druid to fetch additional fields, then outputs an enhanced CSV to standard output.
+# Usage: enhance-report-csv [options] < input.csv > output.csv
+# To use locally, tunnel to solr with: ssh -L 8990:sul-solr-prod-a.stanford.edu:80 lyberadmin@argo-prod-02.stanford.edu
+
+FIELD_HEADERS = {
+  'display_title_ss' => 'title',
+  'member_of_collection_ssim' => 'collection_druids',
+  'collection_title_ssimdv' => 'collection_titles',
+  'governed_by_ssim' => 'apo_druid',
+  'apo_title_ssimdv' => 'apo_title',
+  'folio_instance_hrid_ssim' => 'folio_hrid'
+}.freeze
+
+def parse_options # rubocop:disable Metrics/MethodLength
+  options = {
+    solr_url: 'http://localhost:8990/solr/argo_prod',
+    batch_size: 100
+  }
+
+  OptionParser.new do |opts|
+    opts.banner = 'Usage: enhance-report-csv [options] < input.csv > output.csv'
+
+    opts.on('--solr-url URL', 'Solr URL (default: http://localhost:8990/solr/argo_prod)') do |url|
+      options[:solr_url] = url
+    end
+
+    opts.on('--batch-size NUM', Integer, 'Solr batch size (default: 100)') do |n|
+      options[:batch_size] = n
+    end
+
+    opts.on('-h', '--help', 'Display this help message') do
+      puts opts
+      exit
+    end
+  end.parse!
+
+  options
+end
+
+def fetch_solr_docs(solr, druids)
+  druids.map { |d| "id:(#{d})" }.join(' OR ')
+  response = solr.get('select', params: {
+                        q: '*:*',
+                        fq: "{!terms f=id}#{druids.join(',')}",
+                        fl: "id,#{FIELD_HEADERS.keys.join(',')}",
+                        rows: druids.size
+                      })
+  response['response']['docs'].to_h do |doc|
+    [doc['id'], doc]
+  end
+end
+
+def extract_fields(doc)
+  FIELD_HEADERS.keys.map do |field|
+    value = doc&.fetch(field, nil)
+    value.is_a?(Array) ? value.join(';') : value.to_s
+  end
+end
+
+def build_output(solr, rows, batch_size)
+  extra_col_count = (rows.first&.size || 1) - 1
+  extra_headers = extra_col_count.times.map { |i| "col#{i + 2}" }
+
+  CSV.generate do |out|
+    out << (['druid'] + FIELD_HEADERS.values + extra_headers)
+
+    rows.each_slice(batch_size) do |batch|
+      docs = fetch_solr_docs(solr, batch.map { |row| row[0] })
+
+      batch.each do |row|
+        druid = row[0]
+        out << ([druid] + extract_fields(docs[druid]) + row[1..])
+      end
+    end
+  end
+end
+
+options = parse_options
+solr = RSolr.connect(url: options[:solr_url])
+rows = CSV.parse($stdin.read, headers: false)
+print build_output(solr, rows, options[:batch_size])

data/bin/validate-data

@@ -72,13 +72,18 @@ end
 
 # Get total line count (either from option or by counting)
 def get_total_lines(filename, provided_count)
+  count_filename = filename.sub(/\.xz$/, '.count.txt')
   if provided_count
     puts "Using provided line count: #{provided_count}"
     provided_count
+  elsif File.exist?(count_filename)
+    puts "Reading line count from #{count_filename}..."
+    File.read(count_filename).to_i
   else
     puts 'Counting lines...'
     total = count_lines(filename)
     puts "Total lines to validate: #{total}"
+    File.write(count_filename, total)
     total
   end
 end
@@ -191,6 +196,8 @@ def distribute_work(filename, workers, batch_size, total_lines) # rubocop:disabl
 
       # Update progress bar
       progressbar.increment
+
+      break if line_number >= total_lines
     end
   end
 

data/bin/validate-schema

@@ -5,4 +5,9 @@ require 'bundler/setup'
 require 'cocina/models'
 
 filepath = ARGV[0]
-exit(1) unless JSONSchemer.valid_schema?(Pathname.new(filepath))
+begin
+  JSONSchema.validator_for(JSON.parse(File.read(filepath)))
+rescue StandardError => e
+  warn e.message
+  exit(1)
+end

data/cocina-models.gemspec

@@ -31,8 +31,7 @@ Gem::Specification.new do |spec|
   spec.add_dependency 'edtf' # used for date/time validation
   spec.add_dependency 'equivalent-xml' # for diffing MODS
   spec.add_dependency 'i18n' # for validating BCP 47 language tags, according to RFC 4646
-  spec.add_dependency 'jsonpath' # used for date/time validation
-  spec.add_dependency 'json_schemer', '~> 2.0'
+  spec.add_dependency 'jsonschema_rs'
   spec.add_dependency 'nokogiri'
   spec.add_dependency 'super_diff'
   spec.add_dependency 'thor'

data/lib/cocina/models/contributor.rb

@@ -23,9 +23,6 @@ module Cocina
       attribute :note, Types::Strict::Array.of(DescriptiveValue).default([].freeze)
       # URL or other pointer to the location of the contributor information.
       attribute? :valueAt, Types::Strict::String
-      # For multiple representations of information about the same contributor (e.g. in different
-      # languages).
-      attribute :parallelContributor, Types::Strict::Array.of(DescriptiveParallelContributor).default([].freeze)
     end
   end
 end