@crazygiscool/cap 1.1.10 → 1.2.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (37) hide show
  1. package/dist/__tests__/crawler.test.d.ts +1 -0
  2. package/dist/__tests__/crawler.test.js +175 -0
  3. package/dist/__tests__/crawler.test.js.map +1 -0
  4. package/dist/__tests__/merger.test.d.ts +1 -0
  5. package/dist/__tests__/merger.test.js +136 -0
  6. package/dist/__tests__/merger.test.js.map +1 -0
  7. package/dist/crawler/index.d.ts +11 -0
  8. package/dist/crawler/index.js +119 -0
  9. package/dist/crawler/index.js.map +1 -0
  10. package/dist/crawler/mediawiki.d.ts +18 -0
  11. package/dist/crawler/mediawiki.js +111 -0
  12. package/dist/crawler/mediawiki.js.map +1 -0
  13. package/dist/crawler/merger.d.ts +3 -0
  14. package/dist/crawler/merger.js +128 -0
  15. package/dist/crawler/merger.js.map +1 -0
  16. package/dist/crawler/parser.d.ts +8 -0
  17. package/dist/crawler/parser.js +95 -0
  18. package/dist/crawler/parser.js.map +1 -0
  19. package/dist/crawler/parsers/fandom.d.ts +2 -0
  20. package/dist/crawler/parsers/fandom.js +5 -0
  21. package/dist/crawler/parsers/fandom.js.map +1 -0
  22. package/dist/crawler/parsers/index.d.ts +5 -0
  23. package/dist/crawler/parsers/index.js +17 -0
  24. package/dist/crawler/parsers/index.js.map +1 -0
  25. package/dist/crawler/parsers/tfwiki.d.ts +2 -0
  26. package/dist/crawler/parsers/tfwiki.js +5 -0
  27. package/dist/crawler/parsers/tfwiki.js.map +1 -0
  28. package/dist/crawler/parsers/wikipedia.d.ts +2 -0
  29. package/dist/crawler/parsers/wikipedia.js +5 -0
  30. package/dist/crawler/parsers/wikipedia.js.map +1 -0
  31. package/dist/crawler/types.d.ts +61 -0
  32. package/dist/crawler/types.js +2 -0
  33. package/dist/crawler/types.js.map +1 -0
  34. package/dist/index.d.ts +150 -1
  35. package/dist/index.js +26 -0
  36. package/dist/index.js.map +1 -1
  37. package/package.json +8 -1
@@ -0,0 +1,128 @@
1
+ function sourceExists(sources, url) {
2
+ return (sources ?? []).some((s) => s.url === url);
3
+ }
4
+ export function defaultMergeRules() {
5
+ return {
6
+ "sources-merge": true,
7
+ "longer-description": true,
8
+ "factions-union": true,
9
+ "fill-empty-fields": true,
10
+ "existing-wins": false,
11
+ "crawled-wins": false,
12
+ "continuity-existing": true,
13
+ "bump-updated-at": true,
14
+ "skip-if-no-name": true,
15
+ "skip-if-source-exists": true,
16
+ };
17
+ }
18
+ export function mergeEntry(existing, crawled, rules) {
19
+ const mergedRules = { ...defaultMergeRules(), ...rules };
20
+ const changes = [];
21
+ const doc = {};
22
+ if (mergedRules["skip-if-no-name"] && !crawled.name) {
23
+ return { action: "skipped", changes: ["no name"] };
24
+ }
25
+ if (mergedRules["skip-if-source-exists"] && existing) {
26
+ if (sourceExists(existing.sources, crawled.source.url)) {
27
+ return { action: "skipped", changes: ["source already exists"] };
28
+ }
29
+ }
30
+ if (!existing) {
31
+ // New entry — set everything
32
+ doc.name = crawled.name;
33
+ doc.description = crawled.description;
34
+ doc.continuityId = crawled.continuityId;
35
+ doc.factions = crawled.factions;
36
+ doc.sources = [crawled.source];
37
+ for (const [key, value] of Object.entries(crawled.customFields)) {
38
+ doc[key] = value;
39
+ }
40
+ return { action: "new", doc: doc, changes: ["new entry"] };
41
+ }
42
+ const mergedSources = [...(existing.sources ?? [])];
43
+ if (mergedRules["sources-merge"]) {
44
+ if (!sourceExists(mergedSources, crawled.source.url)) {
45
+ mergedSources.push(crawled.source);
46
+ changes.push("source added");
47
+ }
48
+ }
49
+ if (mergedRules["longer-description"]) {
50
+ if (crawled.description &&
51
+ crawled.description.length > (existing.description ?? "").length) {
52
+ doc.description = crawled.description;
53
+ changes.push("description updated");
54
+ }
55
+ else if ((existing.description ?? "").length > 0) {
56
+ doc.description = existing.description;
57
+ }
58
+ }
59
+ else {
60
+ doc.description = crawled.description || existing.description;
61
+ }
62
+ if (mergedRules["factions-union"]) {
63
+ const all = new Set([
64
+ ...(existing.factions ?? []).map((f) => f.toLowerCase()),
65
+ ...crawled.factions.map((f) => f.toLowerCase()),
66
+ ]);
67
+ const mergedFactions = [...all].map((f) => existing.factions?.find((ef) => ef.toLowerCase() === f) ??
68
+ crawled.factions.find((cf) => cf.toLowerCase() === f));
69
+ // Use original casing from existing if available
70
+ const existingLookup = new Map((existing.factions ?? []).map((f) => [f.toLowerCase(), f]));
71
+ const crawledLookup = new Map(crawled.factions.map((f) => [f.toLowerCase(), f]));
72
+ const union = [];
73
+ for (const lower of all) {
74
+ union.push(existingLookup.get(lower) ?? crawledLookup.get(lower));
75
+ }
76
+ if (union.length !== (existing.factions ?? []).length) {
77
+ changes.push("factions updated");
78
+ }
79
+ doc.factions = union;
80
+ }
81
+ else {
82
+ doc.factions = existing.factions ?? crawled.factions;
83
+ }
84
+ if (mergedRules["continuity-existing"] && existing.continuityId) {
85
+ doc.continuityId = existing.continuityId;
86
+ }
87
+ else {
88
+ doc.continuityId = crawled.continuityId || existing.continuityId;
89
+ }
90
+ // Custom fields
91
+ for (const [key, value] of Object.entries(crawled.customFields)) {
92
+ const existingValue = existing[key];
93
+ if (mergedRules["crawled-wins"]) {
94
+ doc[key] = value;
95
+ }
96
+ else if (mergedRules["existing-wins"]) {
97
+ if (existingValue !== undefined && existingValue !== null) {
98
+ doc[key] = existingValue;
99
+ }
100
+ else {
101
+ doc[key] = value;
102
+ if (value !== undefined && value !== null)
103
+ changes.push(`${key} filled`);
104
+ }
105
+ }
106
+ else if (mergedRules["fill-empty-fields"]) {
107
+ // Default behavior: fill if existing is null/undefined, keep if set
108
+ if (existingValue === undefined || existingValue === null) {
109
+ doc[key] = value;
110
+ if (value !== undefined && value !== null)
111
+ changes.push(`${key} filled`);
112
+ }
113
+ else {
114
+ doc[key] = existingValue;
115
+ }
116
+ }
117
+ }
118
+ doc.sources = mergedSources;
119
+ if (changes.length === 0 && mergedRules["bump-updated-at"]) {
120
+ changes.push("metadata only");
121
+ }
122
+ return {
123
+ action: changes.length > 0 ? "merged" : "skipped",
124
+ doc: doc,
125
+ changes,
126
+ };
127
+ }
128
+ //# sourceMappingURL=merger.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"merger.js","sourceRoot":"","sources":["../../src/crawler/merger.ts"],"names":[],"mappings":"AAQA,SAAS,YAAY,CACnB,OAAwC,EACxC,GAAW;IAEX,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,KAAK,GAAG,CAAC,CAAC;AACpD,CAAC;AAED,MAAM,UAAU,iBAAiB;IAC/B,OAAO;QACL,eAAe,EAAE,IAAI;QACrB,oBAAoB,EAAE,IAAI;QAC1B,gBAAgB,EAAE,IAAI;QACtB,mBAAmB,EAAE,IAAI;QACzB,eAAe,EAAE,KAAK;QACtB,cAAc,EAAE,KAAK;QACrB,qBAAqB,EAAE,IAAI;QAC3B,iBAAiB,EAAE,IAAI;QACvB,iBAAiB,EAAE,IAAI;QACvB,uBAAuB,EAAE,IAAI;KAC9B,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,UAAU,CACxB,QAAkB,EAClB,OAAwB,EACxB,KAAkB;IAElB,MAAM,WAAW,GAAG,EAAE,GAAG,iBAAiB,EAAE,EAAE,GAAG,KAAK,EAAE,CAAC;IACzD,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,MAAM,GAAG,GAA4B,EAAE,CAAC;IAExC,IAAI,WAAW,CAAC,iBAAiB,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC;QACpD,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC,SAAS,CAAC,EAAE,CAAC;IACrD,CAAC;IAED,IAAI,WAAW,CAAC,uBAAuB,CAAC,IAAI,QAAQ,EAAE,CAAC;QACrD,IAAI,YAAY,CAAC,QAAQ,CAAC,OAAO,EAAE,OAAO,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC;YACvD,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC,uBAAuB,CAAC,EAAE,CAAC;QACnE,CAAC;IACH,CAAC;IAED,IAAI,CAAC,QAAQ,EAAE,CAAC;QACd,6BAA6B;QAC7B,GAAG,CAAC,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC;QACxB,GAAG,CAAC,WAAW,GAAG,OAAO,CAAC,WAAW,CAAC;QACtC,GAAG,CAAC,YAAY,GAAG,OAAO,CAAC,YAAY,CAAC;QACxC,GAAG,CAAC,QAAQ,GAAG,OAAO,CAAC,QAAQ,CAAC;QAChC,GAAG,CAAC,OAAO,GAAG,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;QAC/B,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,YAAY,CAAC,EAAE,CAAC;YAChE,GAAG,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC;QACnB,CAAC;QACD,OAAO,EAAE,MAAM,EAAE,KAAK,EAAE,GAAG,EAAE,GAAiB,EAAE,OAAO,EAAE,CAAC,WAAW,CAAC,EAAE,CAAC;IAC3E,CAAC;IAED,MAAM,aAAa,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,CAAC;IAEpD,IAAI,WAAW,CAAC,eAAe,CAAC,EAAE,CAAC;QACjC,IAAI,CAAC,YAAY,CAAC,aAAa,EAAE,OAAO,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC;YACrD,aAAa,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC;YACnC,OAAO,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC;QAC/B,CAAC;IACH,CAAC;IAED,IAAI,WAAW,CAAC,oBAAoB,CAAC,EAAE,CAAC;QACtC,IACE,OAAO,CAAC,WAAW;YACnB,OAAO,CAAC,WAAW,CAAC,MAAM,GAAG,CAAC,QAAQ,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC,MAAM,EAChE,CAAC;YACD,GAAG,CAAC,WAAW,GAAG,OAAO,CAAC,WAAW,CAAC;YACtC,OAAO,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAC;QACtC,CAAC;aAAM,IAAI,CAAC,QAAQ,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACnD,GAAG,CAAC,WAAW,GAAG,QAAQ,CAAC,WAAY,CAAC;QAC1C,CAAC;IACH,CAAC;SAAM,CAAC;QACN,GAAG,CAAC,WAAW,GAAG,OAAO,CAAC,WAAW,IAAI,QAAQ,CAAC,WAAW,CAAC;IAChE,CAAC;IAED,IAAI,WAAW,CAAC,gBAAgB,CAAC,EAAE,CAAC;QAClC,MAAM,GAAG,GAAG,IAAI,GAAG,CAAC;YAClB,GAAG,CAAC,QAAQ,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC;YACxD,GAAG,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC;SAChD,CAAC,CAAC;QACH,MAAM,cAAc,GAAG,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,CACjC,CAAC,CAAC,EAAE,EAAE,CACJ,QAAQ,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,WAAW,EAAE,KAAK,CAAC,CAAC;YACvD,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,WAAW,EAAE,KAAK,CAAC,CAAE,CACzD,CAAC;QACF,iDAAiD;QACjD,MAAM,cAAc,GAAG,IAAI,GAAG,CAC5B,CAAC,QAAQ,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,CAAC,CAC3D,CAAC;QACF,MAAM,aAAa,GAAG,IAAI,GAAG,CAC3B,OAAO,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,WAAW,EAAE,EAAE,CAAC,CAAC,CAAC,CAClD,CAAC;QACF,MAAM,KAAK,GAAa,EAAE,CAAC;QAC3B,KAAK,MAAM,KAAK,IAAI,GAAG,EAAE,CAAC;YACxB,KAAK,CAAC,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,aAAa,CAAC,GAAG,CAAC,KAAK,CAAE,CAAC,CAAC;QACrE,CAAC;QACD,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,QAAQ,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC;YACtD,OAAO,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAAC;QACnC,CAAC;QACD,GAAG,CAAC,QAAQ,GAAG,KAAK,CAAC;IACvB,CAAC;SAAM,CAAC;QACN,GAAG,CAAC,QAAQ,GAAG,QAAQ,CAAC,QAAQ,IAAI,OAAO,CAAC,QAAQ,CAAC;IACvD,CAAC;IAED,IAAI,WAAW,CAAC,qBAAqB,CAAC,IAAI,QAAQ,CAAC,YAAY,EAAE,CAAC;QAChE,GAAG,CAAC,YAAY,GAAG,QAAQ,CAAC,YAAY,CAAC;IAC3C,CAAC;SAAM,CAAC;QACN,GAAG,CAAC,YAAY,GAAG,OAAO,CAAC,YAAY,IAAI,QAAQ,CAAC,YAAY,CAAC;IACnE,CAAC;IAED,gBAAgB;IAChB,KAAK,MAAM,CAAC,GAAG,EAAE,KAAK,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,OAAO,CAAC,YAAY,CAAC,EAAE,CAAC;QAChE,MAAM,aAAa,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC;QACpC,IAAI,WAAW,CAAC,cAAc,CAAC,EAAE,CAAC;YAChC,GAAG,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC;QACnB,CAAC;aAAM,IAAI,WAAW,CAAC,eAAe,CAAC,EAAE,CAAC;YACxC,IAAI,aAAa,KAAK,SAAS,IAAI,aAAa,KAAK,IAAI,EAAE,CAAC;gBAC1D,GAAG,CAAC,GAAG,CAAC,GAAG,aAAa,CAAC;YAC3B,CAAC;iBAAM,CAAC;gBACN,GAAG,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC;gBACjB,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,IAAI;oBAAE,OAAO,CAAC,IAAI,CAAC,GAAG,GAAG,SAAS,CAAC,CAAC;YAC3E,CAAC;QACH,CAAC;aAAM,IAAI,WAAW,CAAC,mBAAmB,CAAC,EAAE,CAAC;YAC5C,oEAAoE;YACpE,IAAI,aAAa,KAAK,SAAS,IAAI,aAAa,KAAK,IAAI,EAAE,CAAC;gBAC1D,GAAG,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC;gBACjB,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,IAAI;oBAAE,OAAO,CAAC,IAAI,CAAC,GAAG,GAAG,SAAS,CAAC,CAAC;YAC3E,CAAC;iBAAM,CAAC;gBACN,GAAG,CAAC,GAAG,CAAC,GAAG,aAAa,CAAC;YAC3B,CAAC;QACH,CAAC;IACH,CAAC;IAED,GAAG,CAAC,OAAO,GAAG,aAAa,CAAC;IAE5B,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,IAAI,WAAW,CAAC,iBAAiB,CAAC,EAAE,CAAC;QAC3D,OAAO,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC;IAChC,CAAC;IAED,OAAO;QACL,MAAM,EAAE,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,SAAS;QACjD,GAAG,EAAE,GAAiB;QACtB,OAAO;KACR,CAAC;AACJ,CAAC"}
@@ -0,0 +1,8 @@
1
+ import * as cheerio from "cheerio";
2
+ import type { ParsedWikiPage } from "./types.js";
3
+ export type { ParsedWikiPage };
4
+ export declare function extractFirstParagraph($: cheerio.CheerioAPI): string;
5
+ export declare function extractCategories($: cheerio.CheerioAPI): string[];
6
+ export declare function extractPageTitle($: cheerio.CheerioAPI): string;
7
+ export declare function parseGenericInfobox(html: string, tableSelector: string, labelSelector: string, valueSelector: string, titleSelector?: string): ParsedWikiPage;
8
+ export declare function parsePortableInfobox(html: string): ParsedWikiPage;
@@ -0,0 +1,95 @@
1
+ import * as cheerio from "cheerio";
2
+ export function extractFirstParagraph($) {
3
+ const content = $("#mw-content-text")[0] ?? $("body")[0];
4
+ if (!content)
5
+ return "";
6
+ const paragraphs = $(content).find("p");
7
+ for (const p of paragraphs) {
8
+ const text = $(p).text().trim();
9
+ if (text.length > 20)
10
+ return text;
11
+ }
12
+ return "";
13
+ }
14
+ export function extractCategories($) {
15
+ const cats = [];
16
+ const catLinks = $("#mw-normal-catlinks ul li a");
17
+ if (catLinks.length) {
18
+ catLinks.each((_, el) => {
19
+ const text = $(el).text().trim();
20
+ if (text)
21
+ cats.push(text);
22
+ });
23
+ }
24
+ return cats;
25
+ }
26
+ export function extractPageTitle($) {
27
+ const h1 = $("#firstHeading").text().trim();
28
+ if (h1)
29
+ return h1;
30
+ return $("title").text().replace(/ – .+$/, "").replace(/ - .+$/, "").trim();
31
+ }
32
+ function cleanValue(value) {
33
+ return value.replace(/<[^>]+>/g, "").replace(/\s+/g, " ").trim();
34
+ }
35
+ export function parseGenericInfobox(html, tableSelector, labelSelector, valueSelector, titleSelector = "") {
36
+ const $ = cheerio.load(html);
37
+ const title = extractPageTitle($);
38
+ const description = extractFirstParagraph($);
39
+ const categories = extractCategories($);
40
+ const infoboxFields = [];
41
+ const table = $(tableSelector).first();
42
+ if (table.length) {
43
+ if (titleSelector) {
44
+ table.find(titleSelector).each((_, el) => {
45
+ const colspan = $(el).attr("colspan");
46
+ if (colspan) {
47
+ const val = $(el).text().trim();
48
+ if (val) {
49
+ infoboxFields.push({ label: "_title", value: val });
50
+ }
51
+ }
52
+ });
53
+ }
54
+ const rows = table.find("tr");
55
+ rows.each((_, row) => {
56
+ const labelEl = $(row).find(labelSelector).first();
57
+ const valueEl = $(row).find(valueSelector).first();
58
+ if (labelEl.length && valueEl.length) {
59
+ const label = labelEl.text().trim();
60
+ const rawHtml = valueEl.html() ?? "";
61
+ const value = cleanValue(rawHtml);
62
+ if (label && label !== "_title" && !label.startsWith("Image")) {
63
+ infoboxFields.push({ label, value });
64
+ }
65
+ }
66
+ });
67
+ }
68
+ return { title, description, categories, infoboxFields };
69
+ }
70
+ export function parsePortableInfobox(html) {
71
+ const $ = cheerio.load(html);
72
+ const title = extractPageTitle($);
73
+ const description = extractFirstParagraph($);
74
+ const categories = extractCategories($);
75
+ const infoboxFields = [];
76
+ const aside = $("aside.portable-infobox").first();
77
+ if (aside.length) {
78
+ const titleEl = aside.find(".pi-title, .pi-data-value.pi-item");
79
+ if (titleEl.length) {
80
+ infoboxFields.push({ label: "_title", value: titleEl.first().text().trim() });
81
+ }
82
+ aside.find('[class*="pi-item"][class*="pi-data"]').each((_, item) => {
83
+ const labelEl = $(item).find("h3");
84
+ const valueEl = $(item).find("div").last();
85
+ if (labelEl.length && valueEl.length) {
86
+ const label = labelEl.text().trim();
87
+ const value = cleanValue(valueEl.html() ?? "");
88
+ if (label)
89
+ infoboxFields.push({ label, value });
90
+ }
91
+ });
92
+ }
93
+ return { title, description, categories, infoboxFields };
94
+ }
95
+ //# sourceMappingURL=parser.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"parser.js","sourceRoot":"","sources":["../../src/crawler/parser.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,OAAO,MAAM,SAAS,CAAC;AAInC,MAAM,UAAU,qBAAqB,CAAC,CAAqB;IACzD,MAAM,OAAO,GAAG,CAAC,CAAC,kBAAkB,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IACzD,IAAI,CAAC,OAAO;QAAE,OAAO,EAAE,CAAC;IAExB,MAAM,UAAU,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACxC,KAAK,MAAM,CAAC,IAAI,UAAU,EAAE,CAAC;QAC3B,MAAM,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;QAChC,IAAI,IAAI,CAAC,MAAM,GAAG,EAAE;YAAE,OAAO,IAAI,CAAC;IACpC,CAAC;IACD,OAAO,EAAE,CAAC;AACZ,CAAC;AAED,MAAM,UAAU,iBAAiB,CAAC,CAAqB;IACrD,MAAM,IAAI,GAAa,EAAE,CAAC;IAC1B,MAAM,QAAQ,GAAG,CAAC,CAAC,6BAA6B,CAAC,CAAC;IAClD,IAAI,QAAQ,CAAC,MAAM,EAAE,CAAC;QACpB,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE;YACtB,MAAM,IAAI,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;YACjC,IAAI,IAAI;gBAAE,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC5B,CAAC,CAAC,CAAC;IACL,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,CAAqB;IACpD,MAAM,EAAE,GAAG,CAAC,CAAC,eAAe,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;IAC5C,IAAI,EAAE;QAAE,OAAO,EAAE,CAAC;IAClB,OAAO,CAAC,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;AAC9E,CAAC;AAED,SAAS,UAAU,CAAC,KAAa;IAC/B,OAAO,KAAK,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AACnE,CAAC;AAED,MAAM,UAAU,mBAAmB,CACjC,IAAY,EACZ,aAAqB,EACrB,aAAqB,EACrB,aAAqB,EACrB,gBAAwB,EAAE;IAE1B,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAE7B,MAAM,KAAK,GAAG,gBAAgB,CAAC,CAAC,CAAC,CAAC;IAClC,MAAM,WAAW,GAAG,qBAAqB,CAAC,CAAC,CAAC,CAAC;IAC7C,MAAM,UAAU,GAAG,iBAAiB,CAAC,CAAC,CAAC,CAAC;IACxC,MAAM,aAAa,GAAuC,EAAE,CAAC;IAE7D,MAAM,KAAK,GAAG,CAAC,CAAC,aAAa,CAAC,CAAC,KAAK,EAAE,CAAC;IACvC,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;QACjB,IAAI,aAAa,EAAE,CAAC;YAClB,KAAK,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE;gBACvC,MAAM,OAAO,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;gBACtC,IAAI,OAAO,EAAE,CAAC;oBACZ,MAAM,GAAG,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;oBAChC,IAAI,GAAG,EAAE,CAAC;wBACR,aAAa,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC;oBACtD,CAAC;gBACH,CAAC;YACH,CAAC,CAAC,CAAC;QACL,CAAC;QAED,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC9B,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,EAAE;YACnB,MAAM,OAAO,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,KAAK,EAAE,CAAC;YACnD,MAAM,OAAO,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,KAAK,EAAE,CAAC;YACnD,IAAI,OAAO,CAAC,MAAM,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;gBACrC,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;gBACpC,MAAM,OAAO,GAAG,OAAO,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC;gBACrC,MAAM,KAAK,GAAG,UAAU,CAAC,OAAO,CAAC,CAAC;gBAClC,IAAI,KAAK,IAAI,KAAK,KAAK,QAAQ,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;oBAC9D,aAAa,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,CAAC;gBACvC,CAAC;YACH,CAAC;QACH,CAAC,CAAC,CAAC;IACL,CAAC;IAED,OAAO,EAAE,KAAK,EAAE,WAAW,EAAE,UAAU,EAAE,aAAa,EAAE,CAAC;AAC3D,CAAC;AAED,MAAM,UAAU,oBAAoB,CAClC,IAAY;IAEZ,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAE7B,MAAM,KAAK,GAAG,gBAAgB,CAAC,CAAC,CAAC,CAAC;IAClC,MAAM,WAAW,GAAG,qBAAqB,CAAC,CAAC,CAAC,CAAC;IAC7C,MAAM,UAAU,GAAG,iBAAiB,CAAC,CAAC,CAAC,CAAC;IACxC,MAAM,aAAa,GAAuC,EAAE,CAAC;IAE7D,MAAM,KAAK,GAAG,CAAC,CAAC,wBAAwB,CAAC,CAAC,KAAK,EAAE,CAAC;IAClD,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;QACjB,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,CAAC,mCAAmC,CAAC,CAAC;QAChE,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;YACnB,aAAa,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;QAChF,CAAC;QAED,KAAK,CAAC,IAAI,CAAC,sCAAsC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,EAAE;YAClE,MAAM,OAAO,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACnC,MAAM,OAAO,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC;YAC3C,IAAI,OAAO,CAAC,MAAM,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;gBACrC,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,CAAC;gBACpC,MAAM,KAAK,GAAG,UAAU,CAAC,OAAO,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC;gBAC/C,IAAI,KAAK;oBAAE,aAAa,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,CAAC;YAClD,CAAC;QACH,CAAC,CAAC,CAAC;IACL,CAAC;IAED,OAAO,EAAE,KAAK,EAAE,WAAW,EAAE,UAAU,EAAE,aAAa,EAAE,CAAC;AAC3D,CAAC"}
@@ -0,0 +1,2 @@
1
+ import { type ParsedWikiPage } from "../parser.js";
2
+ export declare function parseFandom(html: string): ParsedWikiPage;
@@ -0,0 +1,5 @@
1
+ import { parsePortableInfobox } from "../parser.js";
2
+ export function parseFandom(html) {
3
+ return parsePortableInfobox(html);
4
+ }
5
+ //# sourceMappingURL=fandom.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"fandom.js","sourceRoot":"","sources":["../../../src/crawler/parsers/fandom.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,oBAAoB,EAAuB,MAAM,cAAc,CAAC;AAEzE,MAAM,UAAU,WAAW,CAAC,IAAY;IACtC,OAAO,oBAAoB,CAAC,IAAI,CAAC,CAAC;AACpC,CAAC"}
@@ -0,0 +1,5 @@
1
+ export { parseTFWiki } from "./tfwiki.js";
2
+ export { parseWikipedia } from "./wikipedia.js";
3
+ export { parseFandom } from "./fandom.js";
4
+ import type { ParsedWikiPage } from "../parser.js";
5
+ export declare function getParser(name: "tfwiki" | "wikipedia" | "fandom"): (html: string) => ParsedWikiPage;
@@ -0,0 +1,17 @@
1
+ export { parseTFWiki } from "./tfwiki.js";
2
+ export { parseWikipedia } from "./wikipedia.js";
3
+ export { parseFandom } from "./fandom.js";
4
+ import { parseTFWiki } from "./tfwiki.js";
5
+ import { parseWikipedia } from "./wikipedia.js";
6
+ import { parseFandom } from "./fandom.js";
7
+ export function getParser(name) {
8
+ switch (name) {
9
+ case "tfwiki":
10
+ return parseTFWiki;
11
+ case "wikipedia":
12
+ return parseWikipedia;
13
+ case "fandom":
14
+ return parseFandom;
15
+ }
16
+ }
17
+ //# sourceMappingURL=index.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/crawler/parsers/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAC;AAC1C,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAChD,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAC;AAG1C,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAC;AAC1C,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAChD,OAAO,EAAE,WAAW,EAAE,MAAM,aAAa,CAAC;AAE1C,MAAM,UAAU,SAAS,CACvB,IAAuC;IAEvC,QAAQ,IAAI,EAAE,CAAC;QACb,KAAK,QAAQ;YACX,OAAO,WAAW,CAAC;QACrB,KAAK,WAAW;YACd,OAAO,cAAc,CAAC;QACxB,KAAK,QAAQ;YACX,OAAO,WAAW,CAAC;IACvB,CAAC;AACH,CAAC"}
@@ -0,0 +1,2 @@
1
+ import { type ParsedWikiPage } from "../parser.js";
2
+ export declare function parseTFWiki(html: string): ParsedWikiPage;
@@ -0,0 +1,5 @@
1
+ import { parseGenericInfobox } from "../parser.js";
2
+ export function parseTFWiki(html) {
3
+ return parseGenericInfobox(html, "table.infobox", "th", "td", "th[colspan]");
4
+ }
5
+ //# sourceMappingURL=tfwiki.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"tfwiki.js","sourceRoot":"","sources":["../../../src/crawler/parsers/tfwiki.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAuB,MAAM,cAAc,CAAC;AAExE,MAAM,UAAU,WAAW,CAAC,IAAY;IACtC,OAAO,mBAAmB,CACxB,IAAI,EACJ,eAAe,EACf,IAAI,EACJ,IAAI,EACJ,aAAa,CACd,CAAC;AACJ,CAAC"}
@@ -0,0 +1,2 @@
1
+ import { type ParsedWikiPage } from "../parser.js";
2
+ export declare function parseWikipedia(html: string): ParsedWikiPage;
@@ -0,0 +1,5 @@
1
+ import { parseGenericInfobox } from "../parser.js";
2
+ export function parseWikipedia(html) {
3
+ return parseGenericInfobox(html, "table.infobox", "th", "td", "th[colspan]");
4
+ }
5
+ //# sourceMappingURL=wikipedia.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"wikipedia.js","sourceRoot":"","sources":["../../../src/crawler/parsers/wikipedia.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,mBAAmB,EAAuB,MAAM,cAAc,CAAC;AAExE,MAAM,UAAU,cAAc,CAAC,IAAY;IACzC,OAAO,mBAAmB,CACxB,IAAI,EACJ,eAAe,EACf,IAAI,EACJ,IAAI,EACJ,aAAa,CACd,CAAC;AACJ,CAAC"}
@@ -0,0 +1,61 @@
1
+ export interface CrawlSourceConfig {
2
+ id: string;
3
+ label: string;
4
+ baseUrl: string;
5
+ apiUrl: string;
6
+ parser: "tfwiki" | "wikipedia" | "fandom";
7
+ categories: CrawlCategory[];
8
+ fieldMapping: Record<string, FieldMappingEntry>;
9
+ }
10
+ export interface CrawlCategory {
11
+ id: string;
12
+ label: string;
13
+ path: string;
14
+ }
15
+ export interface FieldMappingEntry {
16
+ target: string;
17
+ type: "string" | "number" | "boolean" | "string-array";
18
+ }
19
+ export interface CrawlSourceRecord {
20
+ url: string;
21
+ wiki: string;
22
+ fetchedAt: Date;
23
+ }
24
+ export interface CrawledEntry<T = Record<string, unknown>> {
25
+ name: string;
26
+ description: string;
27
+ continuityId: string;
28
+ factions: string[];
29
+ customFields: Record<string, unknown>;
30
+ source: CrawlSourceRecord;
31
+ }
32
+ export type MergeConfig = Record<string, boolean>;
33
+ export interface MergeResult<T> {
34
+ action: "new" | "merged" | "skipped";
35
+ doc?: Partial<T>;
36
+ changes?: string[];
37
+ }
38
+ export interface MergeableDocument {
39
+ id?: string;
40
+ name: string;
41
+ description?: string;
42
+ continuityId?: string;
43
+ factions?: string[];
44
+ updatedAt?: Date;
45
+ sources?: CrawlSourceRecord[];
46
+ [key: string]: unknown;
47
+ }
48
+ export interface ParsedWikiPage {
49
+ title: string;
50
+ description: string;
51
+ categories: string[];
52
+ infoboxFields: {
53
+ label: string;
54
+ value: string;
55
+ }[];
56
+ }
57
+ export interface CrawlProgress {
58
+ current: number;
59
+ total: number;
60
+ pageTitle: string;
61
+ }
@@ -0,0 +1,2 @@
1
+ export {};
2
+ //# sourceMappingURL=types.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"types.js","sourceRoot":"","sources":["../../src/crawler/types.ts"],"names":[],"mappings":""}
package/dist/index.d.ts CHANGED
@@ -1,5 +1,12 @@
1
1
  import { z } from "zod";
2
- import { Collection, Db } from "mongodb";
2
+ import type { Collection, Db } from "mongodb";
3
+ export { Crawler } from "./crawler/index.js";
4
+ export { MediaWikiClient } from "./crawler/mediawiki.js";
5
+ export { mergeEntry, defaultMergeRules } from "./crawler/merger.js";
6
+ export { getParser, parseTFWiki, parseWikipedia, parseFandom } from "./crawler/parsers/index.js";
7
+ export { parseGenericInfobox, parsePortableInfobox, extractFirstParagraph, extractCategories, extractPageTitle } from "./crawler/parser.js";
8
+ export type { CrawlSourceConfig, CrawlCategory, FieldMappingEntry, CrawlSourceRecord, CrawledEntry, MergeConfig, MergeResult, MergeableDocument, ParsedWikiPage, CrawlProgress, } from "./crawler/types.js";
9
+ import type { CrawlSourceConfig } from "./crawler/types.js";
3
10
  export interface ICAPBaseDocument {
4
11
  id: string;
5
12
  name: string;
@@ -52,9 +59,14 @@ export interface EntryFormatConfig {
52
59
  requiredFields?: string[];
53
60
  customFields?: CustomFieldConfig[];
54
61
  }
62
+ export interface CrawlingConfig {
63
+ sources: CrawlSourceConfig[];
64
+ mergeRules: Record<string, boolean>;
65
+ }
55
66
  export interface SettingsConfig {
56
67
  databaseName: string;
57
68
  entryFormat: EntryFormatConfig;
69
+ crawling?: CrawlingConfig;
58
70
  }
59
71
  export declare const CAPSettingsSchema: z.ZodObject<{
60
72
  databaseName: z.ZodString;
@@ -88,6 +100,105 @@ export declare const CAPSettingsSchema: z.ZodObject<{
88
100
  label: string;
89
101
  }[] | undefined;
90
102
  }>;
103
+ crawling: z.ZodOptional<z.ZodObject<{
104
+ sources: z.ZodArray<z.ZodObject<{
105
+ id: z.ZodString;
106
+ label: z.ZodString;
107
+ baseUrl: z.ZodString;
108
+ apiUrl: z.ZodString;
109
+ parser: z.ZodEnum<["tfwiki", "wikipedia", "fandom"]>;
110
+ categories: z.ZodArray<z.ZodObject<{
111
+ id: z.ZodString;
112
+ label: z.ZodString;
113
+ path: z.ZodString;
114
+ }, "strip", z.ZodTypeAny, {
115
+ id: string;
116
+ path: string;
117
+ label: string;
118
+ }, {
119
+ id: string;
120
+ path: string;
121
+ label: string;
122
+ }>, "many">;
123
+ fieldMapping: z.ZodRecord<z.ZodString, z.ZodObject<{
124
+ target: z.ZodString;
125
+ type: z.ZodEnum<["string", "number", "boolean", "string-array"]>;
126
+ }, "strip", z.ZodTypeAny, {
127
+ type: "string" | "number" | "boolean" | "string-array";
128
+ target: string;
129
+ }, {
130
+ type: "string" | "number" | "boolean" | "string-array";
131
+ target: string;
132
+ }>>;
133
+ }, "strip", z.ZodTypeAny, {
134
+ id: string;
135
+ label: string;
136
+ baseUrl: string;
137
+ apiUrl: string;
138
+ parser: "tfwiki" | "wikipedia" | "fandom";
139
+ categories: {
140
+ id: string;
141
+ path: string;
142
+ label: string;
143
+ }[];
144
+ fieldMapping: Record<string, {
145
+ type: "string" | "number" | "boolean" | "string-array";
146
+ target: string;
147
+ }>;
148
+ }, {
149
+ id: string;
150
+ label: string;
151
+ baseUrl: string;
152
+ apiUrl: string;
153
+ parser: "tfwiki" | "wikipedia" | "fandom";
154
+ categories: {
155
+ id: string;
156
+ path: string;
157
+ label: string;
158
+ }[];
159
+ fieldMapping: Record<string, {
160
+ type: "string" | "number" | "boolean" | "string-array";
161
+ target: string;
162
+ }>;
163
+ }>, "many">;
164
+ mergeRules: z.ZodRecord<z.ZodString, z.ZodBoolean>;
165
+ }, "strip", z.ZodTypeAny, {
166
+ sources: {
167
+ id: string;
168
+ label: string;
169
+ baseUrl: string;
170
+ apiUrl: string;
171
+ parser: "tfwiki" | "wikipedia" | "fandom";
172
+ categories: {
173
+ id: string;
174
+ path: string;
175
+ label: string;
176
+ }[];
177
+ fieldMapping: Record<string, {
178
+ type: "string" | "number" | "boolean" | "string-array";
179
+ target: string;
180
+ }>;
181
+ }[];
182
+ mergeRules: Record<string, boolean>;
183
+ }, {
184
+ sources: {
185
+ id: string;
186
+ label: string;
187
+ baseUrl: string;
188
+ apiUrl: string;
189
+ parser: "tfwiki" | "wikipedia" | "fandom";
190
+ categories: {
191
+ id: string;
192
+ path: string;
193
+ label: string;
194
+ }[];
195
+ fieldMapping: Record<string, {
196
+ type: "string" | "number" | "boolean" | "string-array";
197
+ target: string;
198
+ }>;
199
+ }[];
200
+ mergeRules: Record<string, boolean>;
201
+ }>>;
91
202
  }, "strip", z.ZodTypeAny, {
92
203
  databaseName: string;
93
204
  entryFormat: {
@@ -98,6 +209,25 @@ export declare const CAPSettingsSchema: z.ZodObject<{
98
209
  label: string;
99
210
  }[] | undefined;
100
211
  };
212
+ crawling?: {
213
+ sources: {
214
+ id: string;
215
+ label: string;
216
+ baseUrl: string;
217
+ apiUrl: string;
218
+ parser: "tfwiki" | "wikipedia" | "fandom";
219
+ categories: {
220
+ id: string;
221
+ path: string;
222
+ label: string;
223
+ }[];
224
+ fieldMapping: Record<string, {
225
+ type: "string" | "number" | "boolean" | "string-array";
226
+ target: string;
227
+ }>;
228
+ }[];
229
+ mergeRules: Record<string, boolean>;
230
+ } | undefined;
101
231
  }, {
102
232
  databaseName: string;
103
233
  entryFormat: {
@@ -108,6 +238,25 @@ export declare const CAPSettingsSchema: z.ZodObject<{
108
238
  label: string;
109
239
  }[] | undefined;
110
240
  };
241
+ crawling?: {
242
+ sources: {
243
+ id: string;
244
+ label: string;
245
+ baseUrl: string;
246
+ apiUrl: string;
247
+ parser: "tfwiki" | "wikipedia" | "fandom";
248
+ categories: {
249
+ id: string;
250
+ path: string;
251
+ label: string;
252
+ }[];
253
+ fieldMapping: Record<string, {
254
+ type: "string" | "number" | "boolean" | "string-array";
255
+ target: string;
256
+ }>;
257
+ }[];
258
+ mergeRules: Record<string, boolean>;
259
+ } | undefined;
111
260
  }>;
112
261
  export declare function generateDynamicSchema(customFields: CustomFieldConfig[]): z.ZodObject<z.ZodRawShape>;
113
262
  export declare function loadSettings(path: string): SettingsConfig;
package/dist/index.js CHANGED
@@ -1,6 +1,12 @@
1
1
  import { randomUUID } from "node:crypto";
2
2
  import { readFileSync } from "node:fs";
3
3
  import { z } from "zod";
4
+ // Re-export crawler
5
+ export { Crawler } from "./crawler/index.js";
6
+ export { MediaWikiClient } from "./crawler/mediawiki.js";
7
+ export { mergeEntry, defaultMergeRules } from "./crawler/merger.js";
8
+ export { getParser, parseTFWiki, parseWikipedia, parseFandom } from "./crawler/parsers/index.js";
9
+ export { parseGenericInfobox, parsePortableInfobox, extractFirstParagraph, extractCategories, extractPageTitle } from "./crawler/parser.js";
4
10
  // Runtime validation baseline
5
11
  export const CAPBaseSchema = z.object({
6
12
  name: z.string().min(1),
@@ -48,12 +54,32 @@ const CustomFieldSchema = z.object({
48
54
  type: z.enum(["string", "number", "boolean", "date"]),
49
55
  label: z.string().min(1),
50
56
  });
57
+ const CrawlSourceSchema = z.object({
58
+ id: z.string().min(1),
59
+ label: z.string().min(1),
60
+ baseUrl: z.string().url(),
61
+ apiUrl: z.string().url(),
62
+ parser: z.enum(["tfwiki", "wikipedia", "fandom"]),
63
+ categories: z.array(z.object({
64
+ id: z.string().min(1),
65
+ label: z.string().min(1),
66
+ path: z.string().min(1),
67
+ })),
68
+ fieldMapping: z.record(z.string(), z.object({
69
+ target: z.string().min(1),
70
+ type: z.enum(["string", "number", "boolean", "string-array"]),
71
+ })),
72
+ });
51
73
  export const CAPSettingsSchema = z.object({
52
74
  databaseName: z.string().min(1),
53
75
  entryFormat: z.object({
54
76
  requiredFields: z.array(z.string()).optional(),
55
77
  customFields: z.array(CustomFieldSchema).optional(),
56
78
  }),
79
+ crawling: z.object({
80
+ sources: z.array(CrawlSourceSchema),
81
+ mergeRules: z.record(z.string(), z.boolean()),
82
+ }).optional(),
57
83
  });
58
84
  function zodTypeForFieldType(fieldType) {
59
85
  switch (fieldType) {
package/dist/index.js.map CHANGED
@@ -1 +1 @@
1
- {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,YAAY,EAAE,MAAM,SAAS,CAAC;AACvC,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAqBxB,8BAA8B;AAC9B,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,CAAC,MAAM,CAAC;IACpC,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IACvB,WAAW,EAAE,CAAC,CAAC,MAAM,EAAE;IACvB,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE;IACxB,QAAQ,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC;CAC9B,CAAC,CAAC;AAEH,kCAAkC;AAClC,MAAM,OAAgB,aAAa;IACvB,UAAU,CAAgB;IAEpC,YAAY,EAAM,EAAE,cAAsB;QACxC,IAAI,CAAC,UAAU,GAAG,EAAE,CAAC,UAAU,CAAI,cAAc,CAAC,CAAC;IACrD,CAAC;IAED,KAAK,CAAC,MAAM,CACV,IAA+C;QAE/C,MAAM,GAAG,GAAG;YACV,GAAG,IAAI;YACP,EAAE,EAAE,UAAU,EAAE;YAChB,SAAS,EAAE,IAAI,IAAI,EAAE;YACrB,SAAS,EAAE,IAAI,IAAI,EAAE;SACf,CAAC;QACT,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC;QACrC,OAAO,GAAyB,CAAC;IACnC,CAAC;IAED,KAAK,CAAC,QAAQ,CAAC,EAAU;QACvB,OAAO,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,EAAE,EAAE,EAAS,CAAsB,CAAC;IACrE,CAAC;IAED,KAAK,CAAC,UAAU,CAAC,IAAY;QAC3B,OAAO,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC;YAC7B,IAAI,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,GAAG,EAAE;SAC/B,CAAsB,CAAC;IACjC,CAAC;IAED,KAAK,CAAC,OAAO;QACX,OAAO,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,OAAO,EAAkB,CAAC;IAC5D,CAAC;IAED,KAAK,CAAC,MAAM,CACV,EAAU,EACV,IAAwD;QAExD,OAAO,IAAI,CAAC,UAAU,CAAC,gBAAgB,CACrC,EAAE,EAAE,EAAS,EACb,EAAE,IAAI,EAAE,EAAE,GAAG,IAAI,EAAE,SAAS,EAAE,IAAI,IAAI,EAAE,EAAE,EAAS,EACnD,EAAE,cAAc,EAAE,OAAO,EAAE,CACP,CAAC;IACzB,CAAC;IAED,KAAK,CAAC,MAAM,CAAC,EAAU;QACrB,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,EAAE,EAAE,EAAS,CAAC,CAAC;QAC9D,OAAO,MAAM,CAAC,YAAY,GAAG,CAAC,CAAC;IACjC,CAAC;CACF;AAsBD,MAAM,iBAAiB,GAAG,CAAC,CAAC,MAAM,CAAC;IACjC,GAAG,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IACtB,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,QAAQ,EAAE,SAAS,EAAE,MAAM,CAAC,CAAC;IACrD,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;CACzB,CAAC,CAAC;AAEH,MAAM,CAAC,MAAM,iBAAiB,GAAG,CAAC,CAAC,MAAM,CAAC;IACxC,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IAC/B,WAAW,EAAE,CAAC,CAAC,MAAM,CAAC;QACpB,cAAc,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,QAAQ,EAAE;QAC9C,YAAY,EAAE,CAAC,CAAC,KAAK,CAAC,iBAAiB,CAAC,CAAC,QAAQ,EAAE;KACpD,CAAC;CACH,CAAC,CAAC;AAEH,SAAS,mBAAmB,CAAC,SAAoB;IAC/C,QAAQ,SAAS,EAAE,CAAC;QAClB,KAAK,QAAQ;YACX,OAAO,CAAC,CAAC,MAAM,EAAE,CAAC;QACpB,KAAK,QAAQ;YACX,OAAO,CAAC,CAAC,MAAM,EAAE,CAAC;QACpB,KAAK,SAAS;YACZ,OAAO,CAAC,CAAC,OAAO,EAAE,CAAC;QACrB,KAAK,MAAM;YACT,OAAO,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE,CAAC;QAC/B;YACE,MAAM,IAAI,KAAK,CAAC,uBAAuB,SAAS,EAAE,CAAC,CAAC;IACxD,CAAC;AACH,CAAC;AAED,MAAM,UAAU,qBAAqB,CACnC,YAAiC;IAEjC,MAAM,KAAK,GAA8B,EAAE,CAAC;IAE5C,KAAK,MAAM,KAAK,IAAI,YAAY,EAAE,CAAC;QACjC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,mBAAmB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACrD,CAAC;IAED,OAAO,aAAa,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;AACrC,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,IAAY;IACvC,MAAM,GAAG,GAAG,YAAY,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;IACxC,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAC/B,OAAO,iBAAiB,CAAC,KAAK,CAAC,MAAM,CAAmB,CAAC;AAC3D,CAAC"}
1
+ {"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,YAAY,EAAE,MAAM,SAAS,CAAC;AACvC,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAGxB,oBAAoB;AACpB,OAAO,EAAE,OAAO,EAAE,MAAM,oBAAoB,CAAC;AAC7C,OAAO,EAAE,eAAe,EAAE,MAAM,wBAAwB,CAAC;AACzD,OAAO,EAAE,UAAU,EAAE,iBAAiB,EAAE,MAAM,qBAAqB,CAAC;AACpE,OAAO,EAAE,SAAS,EAAE,WAAW,EAAE,cAAc,EAAE,WAAW,EAAE,MAAM,4BAA4B,CAAC;AACjG,OAAO,EAAE,mBAAmB,EAAE,oBAAoB,EAAE,qBAAqB,EAAE,iBAAiB,EAAE,gBAAgB,EAAE,MAAM,qBAAqB,CAAC;AAiC5I,8BAA8B;AAC9B,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,CAAC,MAAM,CAAC;IACpC,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IACvB,WAAW,EAAE,CAAC,CAAC,MAAM,EAAE;IACvB,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE;IACxB,QAAQ,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC;CAC9B,CAAC,CAAC;AAEH,kCAAkC;AAClC,MAAM,OAAgB,aAAa;IACvB,UAAU,CAAgB;IAEpC,YAAY,EAAM,EAAE,cAAsB;QACxC,IAAI,CAAC,UAAU,GAAG,EAAE,CAAC,UAAU,CAAI,cAAc,CAAC,CAAC;IACrD,CAAC;IAED,KAAK,CAAC,MAAM,CACV,IAA+C;QAE/C,MAAM,GAAG,GAAG;YACV,GAAG,IAAI;YACP,EAAE,EAAE,UAAU,EAAE;YAChB,SAAS,EAAE,IAAI,IAAI,EAAE;YACrB,SAAS,EAAE,IAAI,IAAI,EAAE;SACf,CAAC;QACT,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC;QACrC,OAAO,GAAyB,CAAC;IACnC,CAAC;IAED,KAAK,CAAC,QAAQ,CAAC,EAAU;QACvB,OAAO,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,EAAE,EAAE,EAAS,CAAsB,CAAC;IACrE,CAAC;IAED,KAAK,CAAC,UAAU,CAAC,IAAY;QAC3B,OAAO,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC;YAC7B,IAAI,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,GAAG,EAAE;SAC/B,CAAsB,CAAC;IACjC,CAAC;IAED,KAAK,CAAC,OAAO;QACX,OAAO,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,OAAO,EAAkB,CAAC;IAC5D,CAAC;IAED,KAAK,CAAC,MAAM,CACV,EAAU,EACV,IAAwD;QAExD,OAAO,IAAI,CAAC,UAAU,CAAC,gBAAgB,CACrC,EAAE,EAAE,EAAS,EACb,EAAE,IAAI,EAAE,EAAE,GAAG,IAAI,EAAE,SAAS,EAAE,IAAI,IAAI,EAAE,EAAE,EAAS,EACnD,EAAE,cAAc,EAAE,OAAO,EAAE,CACP,CAAC;IACzB,CAAC;IAED,KAAK,CAAC,MAAM,CAAC,EAAU;QACrB,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,EAAE,EAAE,EAAS,CAAC,CAAC;QAC9D,OAAO,MAAM,CAAC,YAAY,GAAG,CAAC,CAAC;IACjC,CAAC;CACF;AA4BD,MAAM,iBAAiB,GAAG,CAAC,CAAC,MAAM,CAAC;IACjC,GAAG,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IACtB,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,QAAQ,EAAE,SAAS,EAAE,MAAM,CAAC,CAAC;IACrD,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;CACzB,CAAC,CAAC;AAEH,MAAM,iBAAiB,GAAG,CAAC,CAAC,MAAM,CAAC;IACjC,EAAE,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IACrB,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IACxB,OAAO,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE;IACzB,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE;IACxB,MAAM,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,WAAW,EAAE,QAAQ,CAAC,CAAC;IACjD,UAAU,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC;QAC3B,EAAE,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;QACrB,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;QACxB,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;KACxB,CAAC,CAAC;IACH,YAAY,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC;QAC1C,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;QACzB,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,QAAQ,EAAE,SAAS,EAAE,cAAc,CAAC,CAAC;KAC9D,CAAC,CAAC;CACJ,CAAC,CAAC;AAEH,MAAM,CAAC,MAAM,iBAAiB,GAAG,CAAC,CAAC,MAAM,CAAC;IACxC,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IAC/B,WAAW,EAAE,CAAC,CAAC,MAAM,CAAC;QACpB,cAAc,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,QAAQ,EAAE;QAC9C,YAAY,EAAE,CAAC,CAAC,KAAK,CAAC,iBAAiB,CAAC,CAAC,QAAQ,EAAE;KACpD,CAAC;IACF,QAAQ,EAAE,CAAC,CAAC,MAAM,CAAC;QACjB,OAAO,EAAE,CAAC,CAAC,KAAK,CAAC,iBAAiB,CAAC;QACnC,UAAU,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC;KAC9C,CAAC,CAAC,QAAQ,EAAE;CACd,CAAC,CAAC;AAEH,SAAS,mBAAmB,CAAC,SAAoB;IAC/C,QAAQ,SAAS,EAAE,CAAC;QAClB,KAAK,QAAQ;YACX,OAAO,CAAC,CAAC,MAAM,EAAE,CAAC;QACpB,KAAK,QAAQ;YACX,OAAO,CAAC,CAAC,MAAM,EAAE,CAAC;QACpB,KAAK,SAAS;YACZ,OAAO,CAAC,CAAC,OAAO,EAAE,CAAC;QACrB,KAAK,MAAM;YACT,OAAO,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE,CAAC;QAC/B;YACE,MAAM,IAAI,KAAK,CAAC,uBAAuB,SAAS,EAAE,CAAC,CAAC;IACxD,CAAC;AACH,CAAC;AAED,MAAM,UAAU,qBAAqB,CACnC,YAAiC;IAEjC,MAAM,KAAK,GAA8B,EAAE,CAAC;IAE5C,KAAK,MAAM,KAAK,IAAI,YAAY,EAAE,CAAC;QACjC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,GAAG,mBAAmB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACrD,CAAC;IAED,OAAO,aAAa,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;AACrC,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,IAAY;IACvC,MAAM,GAAG,GAAG,YAAY,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;IACxC,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAC/B,OAAO,iBAAiB,CAAC,KAAK,CAAC,MAAM,CAAmB,CAAC;AAC3D,CAAC"}