wikimem 0.2.3 → 0.8.0
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.
- package/CHANGELOG.md +138 -29
- package/README.md +173 -311
- package/dist/cli/commands/ask.d.ts +3 -0
- package/dist/cli/commands/ask.d.ts.map +1 -0
- package/dist/cli/commands/ask.js +63 -0
- package/dist/cli/commands/ask.js.map +1 -0
- package/dist/cli/commands/export.d.ts +3 -0
- package/dist/cli/commands/export.d.ts.map +1 -0
- package/dist/cli/commands/export.js +108 -0
- package/dist/cli/commands/export.js.map +1 -0
- package/dist/cli/commands/history.d.ts +3 -0
- package/dist/cli/commands/history.d.ts.map +1 -0
- package/dist/cli/commands/history.js +61 -0
- package/dist/cli/commands/history.js.map +1 -0
- package/dist/cli/commands/improve.d.ts.map +1 -1
- package/dist/cli/commands/improve.js +4 -3
- package/dist/cli/commands/improve.js.map +1 -1
- package/dist/cli/commands/ingest.d.ts.map +1 -1
- package/dist/cli/commands/ingest.js +5 -4
- package/dist/cli/commands/ingest.js.map +1 -1
- package/dist/cli/commands/init.d.ts.map +1 -1
- package/dist/cli/commands/init.js +246 -79
- package/dist/cli/commands/init.js.map +1 -1
- package/dist/cli/commands/lint.d.ts.map +1 -1
- package/dist/cli/commands/lint.js +4 -3
- package/dist/cli/commands/lint.js.map +1 -1
- package/dist/cli/commands/mcp.d.ts +3 -0
- package/dist/cli/commands/mcp.d.ts.map +1 -0
- package/dist/cli/commands/mcp.js +11 -0
- package/dist/cli/commands/mcp.js.map +1 -0
- package/dist/cli/commands/open.d.ts +3 -0
- package/dist/cli/commands/open.d.ts.map +1 -0
- package/dist/cli/commands/open.js +36 -0
- package/dist/cli/commands/open.js.map +1 -0
- package/dist/cli/commands/query.d.ts.map +1 -1
- package/dist/cli/commands/query.js +5 -4
- package/dist/cli/commands/query.js.map +1 -1
- package/dist/cli/commands/search.d.ts +3 -0
- package/dist/cli/commands/search.d.ts.map +1 -0
- package/dist/cli/commands/search.js +61 -0
- package/dist/cli/commands/search.js.map +1 -0
- package/dist/cli/commands/serve.d.ts.map +1 -1
- package/dist/cli/commands/serve.js +41 -2
- package/dist/cli/commands/serve.js.map +1 -1
- package/dist/cli/commands/watch.d.ts.map +1 -1
- package/dist/cli/commands/watch.js +4 -3
- package/dist/cli/commands/watch.js.map +1 -1
- package/dist/cli/index.d.ts.map +1 -1
- package/dist/cli/index.js +27 -1
- package/dist/cli/index.js.map +1 -1
- package/dist/core/audit-trail.d.ts +15 -0
- package/dist/core/audit-trail.d.ts.map +1 -0
- package/dist/core/audit-trail.js +43 -0
- package/dist/core/audit-trail.js.map +1 -0
- package/dist/core/claude-code.d.ts +10 -0
- package/dist/core/claude-code.d.ts.map +1 -0
- package/dist/core/claude-code.js +81 -0
- package/dist/core/claude-code.js.map +1 -0
- package/dist/core/config.d.ts +23 -0
- package/dist/core/config.d.ts.map +1 -1
- package/dist/core/config.js.map +1 -1
- package/dist/core/connectors.d.ts +58 -0
- package/dist/core/connectors.d.ts.map +1 -0
- package/dist/core/connectors.js +189 -0
- package/dist/core/connectors.js.map +1 -0
- package/dist/core/folder-scanner.d.ts +10 -0
- package/dist/core/folder-scanner.d.ts.map +1 -0
- package/dist/core/folder-scanner.js +84 -0
- package/dist/core/folder-scanner.js.map +1 -0
- package/dist/core/git.d.ts +137 -0
- package/dist/core/git.d.ts.map +1 -0
- package/dist/core/git.js +520 -0
- package/dist/core/git.js.map +1 -0
- package/dist/core/history.d.ts +21 -0
- package/dist/core/history.d.ts.map +1 -0
- package/dist/core/history.js +107 -0
- package/dist/core/history.js.map +1 -0
- package/dist/core/improve.d.ts.map +1 -1
- package/dist/core/improve.js +9 -0
- package/dist/core/improve.js.map +1 -1
- package/dist/core/ingest.d.ts +1 -0
- package/dist/core/ingest.d.ts.map +1 -1
- package/dist/core/ingest.js +78 -5
- package/dist/core/ingest.js.map +1 -1
- package/dist/core/observer.d.ts +71 -0
- package/dist/core/observer.d.ts.map +1 -0
- package/dist/core/observer.js +350 -0
- package/dist/core/observer.js.map +1 -0
- package/dist/core/pipeline-events.d.ts +63 -0
- package/dist/core/pipeline-events.d.ts.map +1 -0
- package/dist/core/pipeline-events.js +109 -0
- package/dist/core/pipeline-events.js.map +1 -0
- package/dist/core/query.d.ts.map +1 -1
- package/dist/core/query.js +16 -8
- package/dist/core/query.js.map +1 -1
- package/dist/core/scraper.d.ts +41 -0
- package/dist/core/scraper.d.ts.map +1 -0
- package/dist/core/scraper.js +277 -0
- package/dist/core/scraper.js.map +1 -0
- package/dist/index.js +3 -2
- package/dist/index.js.map +1 -1
- package/dist/mcp-entry.d.ts +10 -0
- package/dist/mcp-entry.d.ts.map +1 -0
- package/dist/mcp-entry.js +21 -0
- package/dist/mcp-entry.js.map +1 -0
- package/dist/mcp-server.d.ts +15 -0
- package/dist/mcp-server.d.ts.map +1 -0
- package/dist/mcp-server.js +390 -0
- package/dist/mcp-server.js.map +1 -0
- package/dist/processors/audio.d.ts.map +1 -1
- package/dist/processors/audio.js +42 -4
- package/dist/processors/audio.js.map +1 -1
- package/dist/processors/pdf.d.ts.map +1 -1
- package/dist/processors/pdf.js +2 -3
- package/dist/processors/pdf.js.map +1 -1
- package/dist/processors/url.js +4 -1
- package/dist/processors/url.js.map +1 -1
- package/dist/providers/claude.d.ts +1 -0
- package/dist/providers/claude.d.ts.map +1 -1
- package/dist/providers/claude.js +5 -3
- package/dist/providers/claude.js.map +1 -1
- package/dist/providers/index.d.ts +17 -1
- package/dist/providers/index.d.ts.map +1 -1
- package/dist/providers/index.js +144 -0
- package/dist/providers/index.js.map +1 -1
- package/dist/providers/openai.d.ts +1 -0
- package/dist/providers/openai.d.ts.map +1 -1
- package/dist/providers/openai.js +5 -3
- package/dist/providers/openai.js.map +1 -1
- package/dist/providers/types.d.ts +18 -0
- package/dist/providers/types.d.ts.map +1 -1
- package/dist/templates/config-yaml.d.ts.map +1 -1
- package/dist/templates/config-yaml.js +13 -2
- package/dist/templates/config-yaml.js.map +1 -1
- package/dist/web/public/index.html +8157 -745
- package/dist/web/server.d.ts.map +1 -1
- package/dist/web/server.js +2101 -29
- package/dist/web/server.js.map +1 -1
- package/package.json +7 -4
- package/scripts/install.sh +54 -0
- package/src/web/public/index.html +8157 -745
- package/templates/mcp-config.json +9 -0
- package/dist/web/public/public/index.html +0 -946
|
@@ -0,0 +1,41 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Automation 1: Smart Sourcing
|
|
3
|
+
*
|
|
4
|
+
* Full-pipeline scraper that:
|
|
5
|
+
* - Fetches RSS feeds and web pages
|
|
6
|
+
* - Applies topic guardrails (keyword matching)
|
|
7
|
+
* - Labels content as addedBy: 'agent'
|
|
8
|
+
* - Runs ingest pipeline on discovered content
|
|
9
|
+
* - Records every action in the audit trail
|
|
10
|
+
*/
|
|
11
|
+
import type { VaultConfig } from './vault.js';
|
|
12
|
+
import type { UserConfig } from './config.js';
|
|
13
|
+
export interface ScraperOptions {
|
|
14
|
+
/** Skip ingest pipeline — only deposit files to raw/ */
|
|
15
|
+
dryRun?: boolean;
|
|
16
|
+
/** Limit items per source (default: 10) */
|
|
17
|
+
maxItems?: number;
|
|
18
|
+
/** Override topic guardrails for this run */
|
|
19
|
+
skipGuardrails?: boolean;
|
|
20
|
+
}
|
|
21
|
+
export interface ScrapeAction {
|
|
22
|
+
timestamp: string;
|
|
23
|
+
sourceUrl: string;
|
|
24
|
+
sourceName: string;
|
|
25
|
+
itemTitle: string;
|
|
26
|
+
itemUrl: string;
|
|
27
|
+
deposited: boolean;
|
|
28
|
+
ingested: boolean;
|
|
29
|
+
skippedReason?: string;
|
|
30
|
+
}
|
|
31
|
+
export interface SmartScrapeResult {
|
|
32
|
+
sourcesProcessed: number;
|
|
33
|
+
filesDeposited: number;
|
|
34
|
+
pagesCreated: number;
|
|
35
|
+
actions: ScrapeAction[];
|
|
36
|
+
}
|
|
37
|
+
export declare function runSmartScraper(config: VaultConfig, userConfig: UserConfig, options?: ScraperOptions, specificSource?: string): Promise<SmartScrapeResult>;
|
|
38
|
+
export declare function runAdHocScrape(config: VaultConfig, userConfig: UserConfig, urls: string[], options?: ScraperOptions & {
|
|
39
|
+
topics?: string[];
|
|
40
|
+
}): Promise<SmartScrapeResult>;
|
|
41
|
+
//# sourceMappingURL=scraper.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"scraper.d.ts","sourceRoot":"","sources":["../../src/core/scraper.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAIH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AAE9C,OAAO,KAAK,EAAE,UAAU,EAAgB,MAAM,aAAa,CAAC;AAG5D,MAAM,WAAW,cAAc;IAC7B,wDAAwD;IACxD,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB,2CAA2C;IAC3C,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,6CAA6C;IAC7C,cAAc,CAAC,EAAE,OAAO,CAAC;CAC1B;AAED,MAAM,WAAW,YAAY;IAC3B,SAAS,EAAE,MAAM,CAAC;IAClB,SAAS,EAAE,MAAM,CAAC;IAClB,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;IAChB,SAAS,EAAE,OAAO,CAAC;IACnB,QAAQ,EAAE,OAAO,CAAC;IAClB,aAAa,CAAC,EAAE,MAAM,CAAC;CACxB;AAED,MAAM,WAAW,iBAAiB;IAChC,gBAAgB,EAAE,MAAM,CAAC;IACzB,cAAc,EAAE,MAAM,CAAC;IACvB,YAAY,EAAE,MAAM,CAAC;IACrB,OAAO,EAAE,YAAY,EAAE,CAAC;CACzB;AAkGD,wBAAsB,eAAe,CACnC,MAAM,EAAE,WAAW,EACnB,UAAU,EAAE,UAAU,EACtB,OAAO,GAAE,cAAmB,EAC5B,cAAc,CAAC,EAAE,MAAM,GACtB,OAAO,CAAC,iBAAiB,CAAC,CAwC5B;AAED,wBAAsB,cAAc,CAClC,MAAM,EAAE,WAAW,EACnB,UAAU,EAAE,UAAU,EACtB,IAAI,EAAE,MAAM,EAAE,EACd,OAAO,GAAE,cAAc,GAAG;IAAE,MAAM,CAAC,EAAE,MAAM,EAAE,CAAA;CAAO,GACnD,OAAO,CAAC,iBAAiB,CAAC,CAe5B"}
|
|
@@ -0,0 +1,277 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Automation 1: Smart Sourcing
|
|
3
|
+
*
|
|
4
|
+
* Full-pipeline scraper that:
|
|
5
|
+
* - Fetches RSS feeds and web pages
|
|
6
|
+
* - Applies topic guardrails (keyword matching)
|
|
7
|
+
* - Labels content as addedBy: 'agent'
|
|
8
|
+
* - Runs ingest pipeline on discovered content
|
|
9
|
+
* - Records every action in the audit trail
|
|
10
|
+
*/
|
|
11
|
+
import { mkdirSync, writeFileSync } from 'node:fs';
|
|
12
|
+
import { join } from 'node:path';
|
|
13
|
+
import { slugify } from './vault.js';
|
|
14
|
+
import { appendAuditEntry } from './audit-trail.js';
|
|
15
|
+
// ─── Topic Guardrails ────────────────────────────────────────────────────────
|
|
16
|
+
/**
|
|
17
|
+
* Returns true if the content matches the source's topic keywords.
|
|
18
|
+
* If no topics configured, all items pass.
|
|
19
|
+
*/
|
|
20
|
+
function passesGuardrails(title, content, source, skipGuardrails = false) {
|
|
21
|
+
if (skipGuardrails)
|
|
22
|
+
return true;
|
|
23
|
+
const topics = source.topics ?? [];
|
|
24
|
+
if (topics.length === 0)
|
|
25
|
+
return true;
|
|
26
|
+
const haystack = `${title} ${content}`.toLowerCase();
|
|
27
|
+
return topics.some((kw) => haystack.includes(kw.toLowerCase()));
|
|
28
|
+
}
|
|
29
|
+
function parseRssItems(xml) {
|
|
30
|
+
const items = xml.match(/<item[\s>][\s\S]*?<\/item>/g) ?? [];
|
|
31
|
+
return items.map((item) => ({
|
|
32
|
+
title: stripCdata(item.match(/<title>([\s\S]*?)<\/title>/)?.[1] ?? 'Untitled').trim(),
|
|
33
|
+
link: (item.match(/<link>\s*([\s\S]*?)\s*<\/link>/)?.[1] ?? '').trim(),
|
|
34
|
+
description: stripCdata(item.match(/<description>([\s\S]*?)<\/description>/)?.[1] ?? ''),
|
|
35
|
+
pubDate: item.match(/<pubDate>([\s\S]*?)<\/pubDate>/)?.[1]?.trim() ?? '',
|
|
36
|
+
}));
|
|
37
|
+
}
|
|
38
|
+
function parseAtomEntries(xml) {
|
|
39
|
+
const entries = xml.match(/<entry[\s>][\s\S]*?<\/entry>/g) ?? [];
|
|
40
|
+
return entries.map((entry) => ({
|
|
41
|
+
title: stripCdata(entry.match(/<title[^>]*>([\s\S]*?)<\/title>/)?.[1] ?? 'Untitled').trim(),
|
|
42
|
+
link: entry.match(/<link[^>]*href=["']([^"']+)["'][^>]*\/?>/)?.[1]?.trim()
|
|
43
|
+
?? (entry.match(/<link[^>]*>([\s\S]*?)<\/link>/)?.[1] ?? '').trim(),
|
|
44
|
+
description: stripCdata(entry.match(/<summary[^>]*>([\s\S]*?)<\/summary>/)?.[1]
|
|
45
|
+
?? entry.match(/<content[^>]*>([\s\S]*?)<\/content>/)?.[1]
|
|
46
|
+
?? ''),
|
|
47
|
+
pubDate: entry.match(/<(?:updated|published)>([\s\S]*?)<\/(?:updated|published)>/)?.[1]?.trim() ?? '',
|
|
48
|
+
}));
|
|
49
|
+
}
|
|
50
|
+
function parseFeedItems(xml) {
|
|
51
|
+
const rssItems = parseRssItems(xml);
|
|
52
|
+
if (rssItems.length > 0)
|
|
53
|
+
return rssItems;
|
|
54
|
+
return parseAtomEntries(xml);
|
|
55
|
+
}
|
|
56
|
+
async function fetchRssFeed(source, options) {
|
|
57
|
+
if (!source.url)
|
|
58
|
+
return [];
|
|
59
|
+
const res = await fetch(source.url, { signal: AbortSignal.timeout(20000) });
|
|
60
|
+
if (!res.ok)
|
|
61
|
+
throw new Error(`HTTP ${res.status} from ${source.url}`);
|
|
62
|
+
const xml = await res.text();
|
|
63
|
+
const items = parseFeedItems(xml);
|
|
64
|
+
return items.slice(0, options.maxItems ?? 10);
|
|
65
|
+
}
|
|
66
|
+
// ─── Web Scraper ─────────────────────────────────────────────────────────────
|
|
67
|
+
async function fetchPageContent(url) {
|
|
68
|
+
try {
|
|
69
|
+
const res = await fetch(url, {
|
|
70
|
+
signal: AbortSignal.timeout(15000),
|
|
71
|
+
headers: { 'User-Agent': 'WikiMem/0.6 (knowledge-base-bot)' },
|
|
72
|
+
});
|
|
73
|
+
if (!res.ok)
|
|
74
|
+
return '';
|
|
75
|
+
const html = await res.text();
|
|
76
|
+
// Strip scripts/styles first, then all tags
|
|
77
|
+
const cleaned = html
|
|
78
|
+
.replace(/<script[\s\S]*?<\/script>/gi, '')
|
|
79
|
+
.replace(/<style[\s\S]*?<\/style>/gi, '')
|
|
80
|
+
.replace(/<[^>]+>/g, ' ')
|
|
81
|
+
.replace(/&[a-z]+;/gi, ' ')
|
|
82
|
+
.replace(/\s{2,}/g, ' ')
|
|
83
|
+
.trim()
|
|
84
|
+
.substring(0, 8000);
|
|
85
|
+
return cleaned;
|
|
86
|
+
}
|
|
87
|
+
catch {
|
|
88
|
+
return '';
|
|
89
|
+
}
|
|
90
|
+
}
|
|
91
|
+
// ─── Main Scraper Entry Point ─────────────────────────────────────────────────
|
|
92
|
+
export async function runSmartScraper(config, userConfig, options = {}, specificSource) {
|
|
93
|
+
const sources = userConfig.sources ?? [];
|
|
94
|
+
const toProcess = specificSource
|
|
95
|
+
? sources.filter((s) => s.name === specificSource)
|
|
96
|
+
: sources;
|
|
97
|
+
const now = new Date().toISOString().split('T')[0] ?? '';
|
|
98
|
+
const dateDir = join(config.rawDir, now);
|
|
99
|
+
mkdirSync(dateDir, { recursive: true });
|
|
100
|
+
const allActions = [];
|
|
101
|
+
let totalDeposited = 0;
|
|
102
|
+
let totalPagesCreated = 0;
|
|
103
|
+
const startMs = Date.now();
|
|
104
|
+
for (const source of toProcess) {
|
|
105
|
+
const sourceActions = await processSource(source, config, userConfig, dateDir, now, options);
|
|
106
|
+
for (const a of sourceActions.actions) {
|
|
107
|
+
allActions.push(a);
|
|
108
|
+
if (a.deposited)
|
|
109
|
+
totalDeposited++;
|
|
110
|
+
totalPagesCreated += sourceActions.pagesCreated;
|
|
111
|
+
}
|
|
112
|
+
}
|
|
113
|
+
// Record in audit trail
|
|
114
|
+
appendAuditEntry(config.root, {
|
|
115
|
+
action: 'scrape',
|
|
116
|
+
actor: 'agent',
|
|
117
|
+
source: toProcess.map((s) => s.url ?? s.name).join(', '),
|
|
118
|
+
summary: `Smart scraper ran against ${toProcess.length} source(s). Deposited ${totalDeposited} files, created ${totalPagesCreated} pages.`,
|
|
119
|
+
pagesAffected: allActions.filter((a) => a.ingested).map((a) => a.itemTitle),
|
|
120
|
+
duration: Date.now() - startMs,
|
|
121
|
+
});
|
|
122
|
+
return {
|
|
123
|
+
sourcesProcessed: toProcess.length,
|
|
124
|
+
filesDeposited: totalDeposited,
|
|
125
|
+
pagesCreated: totalPagesCreated,
|
|
126
|
+
actions: allActions,
|
|
127
|
+
};
|
|
128
|
+
}
|
|
129
|
+
export async function runAdHocScrape(config, userConfig, urls, options = {}) {
|
|
130
|
+
const sources = urls.map((url) => {
|
|
131
|
+
const isRss = /\.(xml|rss)$/i.test(url) || /\/feed\b|\/rss\b|\/atom\b/i.test(url);
|
|
132
|
+
let hostname = 'unknown';
|
|
133
|
+
try {
|
|
134
|
+
hostname = new URL(url).hostname;
|
|
135
|
+
}
|
|
136
|
+
catch { /* invalid URL */ }
|
|
137
|
+
const source = {
|
|
138
|
+
name: hostname,
|
|
139
|
+
type: isRss ? 'rss' : 'url',
|
|
140
|
+
url,
|
|
141
|
+
};
|
|
142
|
+
if (options.topics)
|
|
143
|
+
source.topics = options.topics;
|
|
144
|
+
return source;
|
|
145
|
+
});
|
|
146
|
+
const tempUserConfig = { ...userConfig, sources };
|
|
147
|
+
return runSmartScraper(config, tempUserConfig, options);
|
|
148
|
+
}
|
|
149
|
+
async function processSource(source, config, userConfig, dateDir, now, options) {
|
|
150
|
+
const actions = [];
|
|
151
|
+
let pagesCreated = 0;
|
|
152
|
+
try {
|
|
153
|
+
if (source.type === 'rss') {
|
|
154
|
+
const items = await fetchRssFeed(source, options);
|
|
155
|
+
for (const item of items) {
|
|
156
|
+
const action = {
|
|
157
|
+
timestamp: new Date().toISOString(),
|
|
158
|
+
sourceUrl: source.url ?? '',
|
|
159
|
+
sourceName: source.name,
|
|
160
|
+
itemTitle: item.title,
|
|
161
|
+
itemUrl: item.link,
|
|
162
|
+
deposited: false,
|
|
163
|
+
ingested: false,
|
|
164
|
+
};
|
|
165
|
+
// Fetch full page content for guardrail check
|
|
166
|
+
const fullContent = item.link ? await fetchPageContent(item.link) : item.description;
|
|
167
|
+
const combinedText = `${item.title} ${item.description} ${fullContent}`;
|
|
168
|
+
if (!passesGuardrails(item.title, combinedText, source, options.skipGuardrails)) {
|
|
169
|
+
action.skippedReason = 'Topic guardrail: no keyword match';
|
|
170
|
+
actions.push(action);
|
|
171
|
+
continue;
|
|
172
|
+
}
|
|
173
|
+
// Build markdown content
|
|
174
|
+
const markdown = buildArticleMarkdown(item.title, item.link, source.name, item.description, fullContent, item.pubDate);
|
|
175
|
+
const fileName = `${slugify(item.title.substring(0, 60))}.md`;
|
|
176
|
+
const filePath = join(dateDir, fileName);
|
|
177
|
+
writeFileSync(filePath, markdown, 'utf-8');
|
|
178
|
+
action.deposited = true;
|
|
179
|
+
// Run ingest pipeline unless dry-run
|
|
180
|
+
if (!options.dryRun) {
|
|
181
|
+
try {
|
|
182
|
+
const result = await ingestFile(filePath, config, userConfig);
|
|
183
|
+
pagesCreated += result.pagesCreated;
|
|
184
|
+
action.ingested = true;
|
|
185
|
+
}
|
|
186
|
+
catch (err) {
|
|
187
|
+
action.skippedReason = `Ingest failed: ${err instanceof Error ? err.message : String(err)}`;
|
|
188
|
+
}
|
|
189
|
+
}
|
|
190
|
+
actions.push(action);
|
|
191
|
+
}
|
|
192
|
+
}
|
|
193
|
+
else if (source.type === 'url') {
|
|
194
|
+
const action = {
|
|
195
|
+
timestamp: new Date().toISOString(),
|
|
196
|
+
sourceUrl: source.url ?? '',
|
|
197
|
+
sourceName: source.name,
|
|
198
|
+
itemTitle: source.name,
|
|
199
|
+
itemUrl: source.url ?? '',
|
|
200
|
+
deposited: false,
|
|
201
|
+
ingested: false,
|
|
202
|
+
};
|
|
203
|
+
const content = await fetchPageContent(source.url ?? '');
|
|
204
|
+
if (!passesGuardrails(source.name, content, source, options.skipGuardrails)) {
|
|
205
|
+
action.skippedReason = 'Topic guardrail: no keyword match';
|
|
206
|
+
actions.push(action);
|
|
207
|
+
}
|
|
208
|
+
else {
|
|
209
|
+
const markdown = `# ${source.name}\n\nSource: ${source.url}\nScraped: ${new Date().toISOString()}\n\n${content}`;
|
|
210
|
+
const fileName = `${slugify(source.name)}.md`;
|
|
211
|
+
const filePath = join(dateDir, fileName);
|
|
212
|
+
writeFileSync(filePath, markdown, 'utf-8');
|
|
213
|
+
action.deposited = true;
|
|
214
|
+
if (!options.dryRun) {
|
|
215
|
+
try {
|
|
216
|
+
const result = await ingestFile(filePath, config, userConfig);
|
|
217
|
+
pagesCreated += result.pagesCreated;
|
|
218
|
+
action.ingested = true;
|
|
219
|
+
}
|
|
220
|
+
catch (err) {
|
|
221
|
+
action.skippedReason = `Ingest failed: ${err instanceof Error ? err.message : String(err)}`;
|
|
222
|
+
}
|
|
223
|
+
}
|
|
224
|
+
actions.push(action);
|
|
225
|
+
}
|
|
226
|
+
}
|
|
227
|
+
}
|
|
228
|
+
catch (err) {
|
|
229
|
+
// Source-level failure — log but don't crash
|
|
230
|
+
const errMsg = err instanceof Error ? err.message : String(err);
|
|
231
|
+
actions.push({
|
|
232
|
+
timestamp: new Date().toISOString(),
|
|
233
|
+
sourceUrl: source.url ?? '',
|
|
234
|
+
sourceName: source.name,
|
|
235
|
+
itemTitle: 'Source fetch failed',
|
|
236
|
+
itemUrl: source.url ?? '',
|
|
237
|
+
deposited: false,
|
|
238
|
+
ingested: false,
|
|
239
|
+
skippedReason: errMsg,
|
|
240
|
+
});
|
|
241
|
+
}
|
|
242
|
+
return { actions, pagesCreated };
|
|
243
|
+
}
|
|
244
|
+
async function ingestFile(filePath, config, userConfig) {
|
|
245
|
+
const { ingestSource } = await import('./ingest.js');
|
|
246
|
+
const { createProviderFromUserConfig } = await import('../providers/index.js');
|
|
247
|
+
const provider = createProviderFromUserConfig(userConfig);
|
|
248
|
+
const result = await ingestSource(filePath, config, provider, {
|
|
249
|
+
verbose: false,
|
|
250
|
+
addedBy: 'agent',
|
|
251
|
+
});
|
|
252
|
+
return { pagesCreated: result.pagesUpdated };
|
|
253
|
+
}
|
|
254
|
+
function buildArticleMarkdown(title, link, sourceName, description, fullContent, pubDate) {
|
|
255
|
+
const parts = [`# ${title}`, ''];
|
|
256
|
+
if (link)
|
|
257
|
+
parts.push(`Source: ${link}`);
|
|
258
|
+
if (sourceName)
|
|
259
|
+
parts.push(`Feed: ${sourceName}`);
|
|
260
|
+
if (pubDate)
|
|
261
|
+
parts.push(`Published: ${pubDate}`);
|
|
262
|
+
parts.push(`Scraped: ${new Date().toISOString()}`, '');
|
|
263
|
+
if (description) {
|
|
264
|
+
parts.push('## Summary', '', stripHtml(description).substring(0, 500), '');
|
|
265
|
+
}
|
|
266
|
+
if (fullContent) {
|
|
267
|
+
parts.push('## Content', '', fullContent.substring(0, 6000));
|
|
268
|
+
}
|
|
269
|
+
return parts.join('\n');
|
|
270
|
+
}
|
|
271
|
+
function stripCdata(text) {
|
|
272
|
+
return text.replace(/<!\[CDATA\[([\s\S]*?)\]\]>/g, '$1').trim();
|
|
273
|
+
}
|
|
274
|
+
function stripHtml(html) {
|
|
275
|
+
return html.replace(/<[^>]*>/g, '').replace(/&[a-z]+;/gi, ' ').trim();
|
|
276
|
+
}
|
|
277
|
+
//# sourceMappingURL=scraper.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"scraper.js","sourceRoot":"","sources":["../../src/core/scraper.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EAAE,SAAS,EAAE,aAAa,EAAE,MAAM,SAAS,CAAC;AACnD,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAEjC,OAAO,EAAE,OAAO,EAAE,MAAM,YAAY,CAAC;AAErC,OAAO,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAC;AA6BpD,gFAAgF;AAEhF;;;GAGG;AACH,SAAS,gBAAgB,CACvB,KAAa,EACb,OAAe,EACf,MAAoB,EACpB,cAAc,GAAG,KAAK;IAEtB,IAAI,cAAc;QAAE,OAAO,IAAI,CAAC;IAChC,MAAM,MAAM,GAAc,MAA+C,CAAC,MAAM,IAAI,EAAE,CAAC;IACvF,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IAErC,MAAM,QAAQ,GAAG,GAAG,KAAK,IAAI,OAAO,EAAE,CAAC,WAAW,EAAE,CAAC;IACrD,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,QAAQ,CAAC,QAAQ,CAAC,EAAE,CAAC,WAAW,EAAE,CAAC,CAAC,CAAC;AAClE,CAAC;AAWD,SAAS,aAAa,CAAC,GAAW;IAChC,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC,6BAA6B,CAAC,IAAI,EAAE,CAAC;IAC7D,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;QAC1B,KAAK,EAAE,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,4BAA4B,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,UAAU,CAAC,CAAC,IAAI,EAAE;QACrF,IAAI,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,gCAAgC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE;QACtE,WAAW,EAAE,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,wCAAwC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QACxF,OAAO,EAAE,IAAI,CAAC,KAAK,CAAC,gCAAgC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE;KACzE,CAAC,CAAC,CAAC;AACN,CAAC;AAED,SAAS,gBAAgB,CAAC,GAAW;IACnC,MAAM,OAAO,GAAG,GAAG,CAAC,KAAK,CAAC,+BAA+B,CAAC,IAAI,EAAE,CAAC;IACjE,OAAO,OAAO,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QAC7B,KAAK,EAAE,UAAU,CAAC,KAAK,CAAC,KAAK,CAAC,iCAAiC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,UAAU,CAAC,CAAC,IAAI,EAAE;QAC3F,IAAI,EAAE,KAAK,CAAC,KAAK,CAAC,0CAA0C,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE;eACrE,CAAC,KAAK,CAAC,KAAK,CAAC,+BAA+B,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE;QACrE,WAAW,EAAE,UAAU,CACrB,KAAK,CAAC,KAAK,CAAC,qCAAqC,CAAC,EAAE,CAAC,CAAC,CAAC;eAClD,KAAK,CAAC,KAAK,CAAC,qCAAqC,CAAC,EAAE,CAAC,CAAC,CAAC;eACvD,EAAE,CACR;QACD,OAAO,EAAE,KAAK,CAAC,KAAK,CAAC,4DAA4D,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE;KACtG,CAAC,CAAC,CAAC;AACN,CAAC;AAED,SAAS,cAAc,CAAC,GAAW;IACjC,MAAM,QAAQ,GAAG,aAAa,CAAC,GAAG,CAAC,CAAC;IACpC,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,QAAQ,CAAC;IACzC,OAAO,gBAAgB,CAAC,GAAG,CAAC,CAAC;AAC/B,CAAC;AAED,KAAK,UAAU,YAAY,CAAC,MAAoB,EAAE,OAAuB;IACvE,IAAI,CAAC,MAAM,CAAC,GAAG;QAAE,OAAO,EAAE,CAAC;IAC3B,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,MAAM,CAAC,GAAG,EAAE,EAAE,MAAM,EAAE,WAAW,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IAC5E,IAAI,CAAC,GAAG,CAAC,EAAE;QAAE,MAAM,IAAI,KAAK,CAAC,QAAQ,GAAG,CAAC,MAAM,SAAS,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC;IACtE,MAAM,GAAG,GAAG,MAAM,GAAG,CAAC,IAAI,EAAE,CAAC;IAC7B,MAAM,KAAK,GAAG,cAAc,CAAC,GAAG,CAAC,CAAC;IAClC,OAAO,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC;AAChD,CAAC;AAED,gFAAgF;AAEhF,KAAK,UAAU,gBAAgB,CAAC,GAAW;IACzC,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,GAAG,EAAE;YAC3B,MAAM,EAAE,WAAW,CAAC,OAAO,CAAC,KAAK,CAAC;YAClC,OAAO,EAAE,EAAE,YAAY,EAAE,kCAAkC,EAAE;SAC9D,CAAC,CAAC;QACH,IAAI,CAAC,GAAG,CAAC,EAAE;YAAE,OAAO,EAAE,CAAC;QACvB,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,IAAI,EAAE,CAAC;QAC9B,4CAA4C;QAC5C,MAAM,OAAO,GAAG,IAAI;aACjB,OAAO,CAAC,6BAA6B,EAAE,EAAE,CAAC;aAC1C,OAAO,CAAC,2BAA2B,EAAE,EAAE,CAAC;aACxC,OAAO,CAAC,UAAU,EAAE,GAAG,CAAC;aACxB,OAAO,CAAC,YAAY,EAAE,GAAG,CAAC;aAC1B,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC;aACvB,IAAI,EAAE;aACN,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;QACtB,OAAO,OAAO,CAAC;IACjB,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,CAAC;IACZ,CAAC;AACH,CAAC;AAED,iFAAiF;AAEjF,MAAM,CAAC,KAAK,UAAU,eAAe,CACnC,MAAmB,EACnB,UAAsB,EACtB,UAA0B,EAAE,EAC5B,cAAuB;IAEvB,MAAM,OAAO,GAAG,UAAU,CAAC,OAAO,IAAI,EAAE,CAAC;IACzC,MAAM,SAAS,GAAG,cAAc;QAC9B,CAAC,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,cAAc,CAAC;QAClD,CAAC,CAAC,OAAO,CAAC;IAEZ,MAAM,GAAG,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;IACzD,MAAM,OAAO,GAAG,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;IACzC,SAAS,CAAC,OAAO,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IAExC,MAAM,UAAU,GAAmB,EAAE,CAAC;IACtC,IAAI,cAAc,GAAG,CAAC,CAAC;IACvB,IAAI,iBAAiB,GAAG,CAAC,CAAC;IAC1B,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAE3B,KAAK,MAAM,MAAM,IAAI,SAAS,EAAE,CAAC;QAC/B,MAAM,aAAa,GAAG,MAAM,aAAa,CAAC,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,OAAO,EAAE,GAAG,EAAE,OAAO,CAAC,CAAC;QAC7F,KAAK,MAAM,CAAC,IAAI,aAAa,CAAC,OAAO,EAAE,CAAC;YACtC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACnB,IAAI,CAAC,CAAC,SAAS;gBAAE,cAAc,EAAE,CAAC;YAClC,iBAAiB,IAAI,aAAa,CAAC,YAAY,CAAC;QAClD,CAAC;IACH,CAAC;IAED,wBAAwB;IACxB,gBAAgB,CAAC,MAAM,CAAC,IAAI,EAAE;QAC5B,MAAM,EAAE,QAAQ;QAChB,KAAK,EAAE,OAAO;QACd,MAAM,EAAE,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC;QACxD,OAAO,EAAE,6BAA6B,SAAS,CAAC,MAAM,yBAAyB,cAAc,mBAAmB,iBAAiB,SAAS;QAC1I,aAAa,EAAE,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC;QAC3E,QAAQ,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,OAAO;KAC/B,CAAC,CAAC;IAEH,OAAO;QACL,gBAAgB,EAAE,SAAS,CAAC,MAAM;QAClC,cAAc,EAAE,cAAc;QAC9B,YAAY,EAAE,iBAAiB;QAC/B,OAAO,EAAE,UAAU;KACpB,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,cAAc,CAClC,MAAmB,EACnB,UAAsB,EACtB,IAAc,EACd,UAAkD,EAAE;IAEpD,MAAM,OAAO,GAAmB,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE;QAC/C,MAAM,KAAK,GAAG,eAAe,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,4BAA4B,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAClF,IAAI,QAAQ,GAAG,SAAS,CAAC;QACzB,IAAI,CAAC;YAAC,QAAQ,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC,QAAQ,CAAC;QAAC,CAAC;QAAC,MAAM,CAAC,CAAC,iBAAiB,CAAC,CAAC;QACrE,MAAM,MAAM,GAAyC;YACnD,IAAI,EAAE,QAAQ;YACd,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK;YAC3B,GAAG;SACJ,CAAC;QACF,IAAI,OAAO,CAAC,MAAM;YAAE,MAAM,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC;QACnD,OAAO,MAAsB,CAAC;IAChC,CAAC,CAAC,CAAC;IACH,MAAM,cAAc,GAAe,EAAE,GAAG,UAAU,EAAE,OAAO,EAAE,CAAC;IAC9D,OAAO,eAAe,CAAC,MAAM,EAAE,cAAc,EAAE,OAAO,CAAC,CAAC;AAC1D,CAAC;AAOD,KAAK,UAAU,aAAa,CAC1B,MAAoB,EACpB,MAAmB,EACnB,UAAsB,EACtB,OAAe,EACf,GAAW,EACX,OAAuB;IAEvB,MAAM,OAAO,GAAmB,EAAE,CAAC;IACnC,IAAI,YAAY,GAAG,CAAC,CAAC;IAErB,IAAI,CAAC;QACH,IAAI,MAAM,CAAC,IAAI,KAAK,KAAK,EAAE,CAAC;YAC1B,MAAM,KAAK,GAAG,MAAM,YAAY,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;YAElD,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;gBACzB,MAAM,MAAM,GAAiB;oBAC3B,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;oBACnC,SAAS,EAAE,MAAM,CAAC,GAAG,IAAI,EAAE;oBAC3B,UAAU,EAAE,MAAM,CAAC,IAAI;oBACvB,SAAS,EAAE,IAAI,CAAC,KAAK;oBACrB,OAAO,EAAE,IAAI,CAAC,IAAI;oBAClB,SAAS,EAAE,KAAK;oBAChB,QAAQ,EAAE,KAAK;iBAChB,CAAC;gBAEF,8CAA8C;gBAC9C,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,WAAW,CAAC;gBACrF,MAAM,YAAY,GAAG,GAAG,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,WAAW,IAAI,WAAW,EAAE,CAAC;gBAExE,IAAI,CAAC,gBAAgB,CAAC,IAAI,CAAC,KAAK,EAAE,YAAY,EAAE,MAAM,EAAE,OAAO,CAAC,cAAc,CAAC,EAAE,CAAC;oBAChF,MAAM,CAAC,aAAa,GAAG,mCAAmC,CAAC;oBAC3D,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;oBACrB,SAAS;gBACX,CAAC;gBAED,yBAAyB;gBACzB,MAAM,QAAQ,GAAG,oBAAoB,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,IAAI,EAAE,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,WAAW,EAAE,WAAW,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC;gBACvH,MAAM,QAAQ,GAAG,GAAG,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC;gBAC9D,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;gBACzC,aAAa,CAAC,QAAQ,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC;gBAC3C,MAAM,CAAC,SAAS,GAAG,IAAI,CAAC;gBAExB,qCAAqC;gBACrC,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC;oBACpB,IAAI,CAAC;wBACH,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC,QAAQ,EAAE,MAAM,EAAE,UAAU,CAAC,CAAC;wBAC9D,YAAY,IAAI,MAAM,CAAC,YAAY,CAAC;wBACpC,MAAM,CAAC,QAAQ,GAAG,IAAI,CAAC;oBACzB,CAAC;oBAAC,OAAO,GAAG,EAAE,CAAC;wBACb,MAAM,CAAC,aAAa,GAAG,kBAAkB,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC;oBAC9F,CAAC;gBACH,CAAC;gBAED,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACvB,CAAC;QACH,CAAC;aAAM,IAAI,MAAM,CAAC,IAAI,KAAK,KAAK,EAAE,CAAC;YACjC,MAAM,MAAM,GAAiB;gBAC3B,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;gBACnC,SAAS,EAAE,MAAM,CAAC,GAAG,IAAI,EAAE;gBAC3B,UAAU,EAAE,MAAM,CAAC,IAAI;gBACvB,SAAS,EAAE,MAAM,CAAC,IAAI;gBACtB,OAAO,EAAE,MAAM,CAAC,GAAG,IAAI,EAAE;gBACzB,SAAS,EAAE,KAAK;gBAChB,QAAQ,EAAE,KAAK;aAChB,CAAC;YAEF,MAAM,OAAO,GAAG,MAAM,gBAAgB,CAAC,MAAM,CAAC,GAAG,IAAI,EAAE,CAAC,CAAC;YACzD,IAAI,CAAC,gBAAgB,CAAC,MAAM,CAAC,IAAI,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO,CAAC,cAAc,CAAC,EAAE,CAAC;gBAC5E,MAAM,CAAC,aAAa,GAAG,mCAAmC,CAAC;gBAC3D,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACvB,CAAC;iBAAM,CAAC;gBACN,MAAM,QAAQ,GAAG,KAAK,MAAM,CAAC,IAAI,eAAe,MAAM,CAAC,GAAG,cAAc,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,OAAO,OAAO,EAAE,CAAC;gBACjH,MAAM,QAAQ,GAAG,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC;gBAC9C,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;gBACzC,aAAa,CAAC,QAAQ,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC;gBAC3C,MAAM,CAAC,SAAS,GAAG,IAAI,CAAC;gBAExB,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,CAAC;oBACpB,IAAI,CAAC;wBACH,MAAM,MAAM,GAAG,MAAM,UAAU,CAAC,QAAQ,EAAE,MAAM,EAAE,UAAU,CAAC,CAAC;wBAC9D,YAAY,IAAI,MAAM,CAAC,YAAY,CAAC;wBACpC,MAAM,CAAC,QAAQ,GAAG,IAAI,CAAC;oBACzB,CAAC;oBAAC,OAAO,GAAG,EAAE,CAAC;wBACb,MAAM,CAAC,aAAa,GAAG,kBAAkB,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC;oBAC9F,CAAC;gBACH,CAAC;gBAED,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YACvB,CAAC;QACH,CAAC;IACH,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,6CAA6C;QAC7C,MAAM,MAAM,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QAChE,OAAO,CAAC,IAAI,CAAC;YACX,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;YACnC,SAAS,EAAE,MAAM,CAAC,GAAG,IAAI,EAAE;YAC3B,UAAU,EAAE,MAAM,CAAC,IAAI;YACvB,SAAS,EAAE,qBAAqB;YAChC,OAAO,EAAE,MAAM,CAAC,GAAG,IAAI,EAAE;YACzB,SAAS,EAAE,KAAK;YAChB,QAAQ,EAAE,KAAK;YACf,aAAa,EAAE,MAAM;SACtB,CAAC,CAAC;IACL,CAAC;IAED,OAAO,EAAE,OAAO,EAAE,YAAY,EAAE,CAAC;AACnC,CAAC;AAED,KAAK,UAAU,UAAU,CACvB,QAAgB,EAChB,MAAmB,EACnB,UAAsB;IAEtB,MAAM,EAAE,YAAY,EAAE,GAAG,MAAM,MAAM,CAAC,aAAa,CAAC,CAAC;IACrD,MAAM,EAAE,4BAA4B,EAAE,GAAG,MAAM,MAAM,CAAC,uBAAuB,CAAC,CAAC;IAC/E,MAAM,QAAQ,GAAG,4BAA4B,CAAC,UAAU,CAAC,CAAC;IAC1D,MAAM,MAAM,GAAG,MAAM,YAAY,CAAC,QAAQ,EAAE,MAAM,EAAE,QAAQ,EAAE;QAC5D,OAAO,EAAE,KAAK;QACd,OAAO,EAAE,OAAO;KACjB,CAAC,CAAC;IACH,OAAO,EAAE,YAAY,EAAE,MAAM,CAAC,YAAY,EAAE,CAAC;AAC/C,CAAC;AAED,SAAS,oBAAoB,CAC3B,KAAa,EACb,IAAY,EACZ,UAAkB,EAClB,WAAmB,EACnB,WAAmB,EACnB,OAAe;IAEf,MAAM,KAAK,GAAa,CAAC,KAAK,KAAK,EAAE,EAAE,EAAE,CAAC,CAAC;IAC3C,IAAI,IAAI;QAAE,KAAK,CAAC,IAAI,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC;IACxC,IAAI,UAAU;QAAE,KAAK,CAAC,IAAI,CAAC,SAAS,UAAU,EAAE,CAAC,CAAC;IAClD,IAAI,OAAO;QAAE,KAAK,CAAC,IAAI,CAAC,cAAc,OAAO,EAAE,CAAC,CAAC;IACjD,KAAK,CAAC,IAAI,CAAC,YAAY,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC;IACvD,IAAI,WAAW,EAAE,CAAC;QAChB,KAAK,CAAC,IAAI,CAAC,YAAY,EAAE,EAAE,EAAE,SAAS,CAAC,WAAW,CAAC,CAAC,SAAS,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,EAAE,CAAC,CAAC;IAC7E,CAAC;IACD,IAAI,WAAW,EAAE,CAAC;QAChB,KAAK,CAAC,IAAI,CAAC,YAAY,EAAE,EAAE,EAAE,WAAW,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC;IAC/D,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1B,CAAC;AAED,SAAS,UAAU,CAAC,IAAY;IAC9B,OAAO,IAAI,CAAC,OAAO,CAAC,6BAA6B,EAAE,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;AAClE,CAAC;AAED,SAAS,SAAS,CAAC,IAAY;IAC7B,OAAO,IAAI,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,YAAY,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AACxE,CAAC"}
|
package/dist/index.js
CHANGED
|
@@ -29,10 +29,11 @@ if (process.argv.length <= 2) {
|
|
|
29
29
|
console.log(chalk.bold(' Commands'));
|
|
30
30
|
console.log(chalk.dim(' ─'.padEnd(40, '─')));
|
|
31
31
|
console.log(` ${chalk.cyan('wikimem ingest <file|url>')} Add knowledge`);
|
|
32
|
-
console.log(` ${chalk.cyan('wikimem
|
|
32
|
+
console.log(` ${chalk.cyan('wikimem search "term"')} BM25 keyword search`);
|
|
33
|
+
console.log(` ${chalk.cyan('wikimem ask "question"')} LLM-powered Q&A`);
|
|
33
34
|
console.log(` ${chalk.cyan('wikimem status')} Full statistics`);
|
|
34
35
|
console.log(` ${chalk.cyan('wikimem lint')} Check wiki health`);
|
|
35
|
-
console.log(` ${chalk.cyan('wikimem
|
|
36
|
+
console.log(` ${chalk.cyan('wikimem export -f json')} Export graph data`);
|
|
36
37
|
console.log(` ${chalk.cyan('wikimem serve')} Web UI at :3141`);
|
|
37
38
|
console.log(` ${chalk.cyan('wikimem --help')} All commands`);
|
|
38
39
|
console.log();
|
package/dist/index.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";AAEA,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AACrC,OAAO,EAAQ,OAAO,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,eAAe,EAAE,MAAM,0BAA0B,CAAC;AAC3D,OAAO,EAAE,cAAc,EAAE,aAAa,EAAE,MAAM,iBAAiB,CAAC;AAEhE,MAAM,OAAO,GAAG,aAAa,EAAE,CAAC;AAEhC,8EAA8E;AAC9E,IAAI,OAAO,CAAC,IAAI,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;IAC7B,MAAM,GAAG,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC;IACzB,MAAM,MAAM,GAAG,cAAc,CAAC,GAAG,CAAC,CAAC;IAEnC,IAAI,UAAU,CAAC,MAAM,CAAC,UAAU,CAAC,EAAE,CAAC;QAClC,iDAAiD;QACjD,MAAM,KAAK,GAAG,aAAa,CAAC,MAAM,CAAC,CAAC;QAEpC,OAAO,CAAC,GAAG,EAAE,CAAC;QACd,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,oBAAoB,CAAC,CAAC,CAAC;QAC7D,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAC,CAAC;QAC9D,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAC,CAAC;QAC9D,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,iCAAiC,CAAC,CAAC,CAAC;QAC1D,OAAO,CAAC,GAAG,EAAE,CAAC;QACd,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAAC,CAAC;QAC5C,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;QAC9C,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,UAAU,KAAK,CAAC,SAAS,EAAE,CAAC,CAAC;QAClE,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,UAAU,KAAK,CAAC,SAAS,CAAC,cAAc,EAAE,EAAE,CAAC,CAAC;QACnF,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;QACpE,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,aAAa,CAAC,KAAK,KAAK,CAAC,SAAS,EAAE,CAAC,CAAC;QAClE,OAAO,CAAC,GAAG,EAAE,CAAC;QACd,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC,CAAC;QACtC,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;QAC9C,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,2BAA2B,CAAC,iBAAiB,CAAC,CAAC;QAC3E,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":";AAEA,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AACrC,OAAO,EAAQ,OAAO,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,EAAE,aAAa,EAAE,MAAM,gBAAgB,CAAC;AAC/C,OAAO,EAAE,eAAe,EAAE,MAAM,0BAA0B,CAAC;AAC3D,OAAO,EAAE,cAAc,EAAE,aAAa,EAAE,MAAM,iBAAiB,CAAC;AAEhE,MAAM,OAAO,GAAG,aAAa,EAAE,CAAC;AAEhC,8EAA8E;AAC9E,IAAI,OAAO,CAAC,IAAI,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;IAC7B,MAAM,GAAG,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC;IACzB,MAAM,MAAM,GAAG,cAAc,CAAC,GAAG,CAAC,CAAC;IAEnC,IAAI,UAAU,CAAC,MAAM,CAAC,UAAU,CAAC,EAAE,CAAC;QAClC,iDAAiD;QACjD,MAAM,KAAK,GAAG,aAAa,CAAC,MAAM,CAAC,CAAC;QAEpC,OAAO,CAAC,GAAG,EAAE,CAAC;QACd,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,oBAAoB,CAAC,CAAC,CAAC;QAC7D,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAC,CAAC;QAC9D,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAC,CAAC;QAC9D,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,iCAAiC,CAAC,CAAC,CAAC;QAC1D,OAAO,CAAC,GAAG,EAAE,CAAC;QACd,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAAC,CAAC;QAC5C,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;QAC9C,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,UAAU,KAAK,CAAC,SAAS,EAAE,CAAC,CAAC;QAClE,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,UAAU,KAAK,CAAC,SAAS,CAAC,cAAc,EAAE,EAAE,CAAC,CAAC;QACnF,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;QACpE,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,aAAa,CAAC,KAAK,KAAK,CAAC,SAAS,EAAE,CAAC,CAAC;QAClE,OAAO,CAAC,GAAG,EAAE,CAAC;QACd,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC,CAAC;QACtC,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;QAC9C,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,2BAA2B,CAAC,iBAAiB,CAAC,CAAC;QAC3E,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,uBAAuB,CAAC,2BAA2B,CAAC,CAAC;QACjF,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,wBAAwB,CAAC,sBAAsB,CAAC,CAAC;QAC7E,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,gBAAgB,CAAC,8BAA8B,CAAC,CAAC;QAC7E,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,cAAc,CAAC,kCAAkC,CAAC,CAAC;QAC/E,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,wBAAwB,CAAC,wBAAwB,CAAC,CAAC;QAC/E,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,eAAe,CAAC,+BAA+B,CAAC,CAAC;QAC7E,OAAO,CAAC,GAAG,CAAC,KAAK,KAAK,CAAC,IAAI,CAAC,gBAAgB,CAAC,2BAA2B,CAAC,CAAC;QAC1E,OAAO,CAAC,GAAG,EAAE,CAAC;IAChB,CAAC;SAAM,CAAC;QACN,6BAA6B;QAC7B,OAAO,CAAC,UAAU,EAAE,CAAC;QACrB,OAAO,CAAC,GAAG,EAAE,CAAC;QACd,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,2BAA2B,CAAC,CAAC,CAAC;IACtD,CAAC;IAED,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAClB,CAAC;KAAM,CAAC;IACN,OAAO,CAAC,KAAK,EAAE,CAAC;AAClB,CAAC;AAED,+DAA+D;AAC/D,eAAe,EAAE,CAAC,KAAK,CAAC,GAAG,EAAE,GAAE,CAAC,CAAC,CAAC"}
|
|
@@ -0,0 +1,10 @@
|
|
|
1
|
+
#!/usr/bin/env node
|
|
2
|
+
/**
|
|
3
|
+
* Lightweight MCP server entry point.
|
|
4
|
+
* Bypasses the heavy CLI import chain (pdf-parse, mammoth, xlsx, etc.)
|
|
5
|
+
* so the server starts in <2s instead of >90s.
|
|
6
|
+
*
|
|
7
|
+
* Usage: node dist/mcp-entry.js [--vault <path>]
|
|
8
|
+
*/
|
|
9
|
+
export {};
|
|
10
|
+
//# sourceMappingURL=mcp-entry.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"mcp-entry.d.ts","sourceRoot":"","sources":["../src/mcp-entry.ts"],"names":[],"mappings":";AAEA;;;;;;GAMG"}
|
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
#!/usr/bin/env node
|
|
2
|
+
/**
|
|
3
|
+
* Lightweight MCP server entry point.
|
|
4
|
+
* Bypasses the heavy CLI import chain (pdf-parse, mammoth, xlsx, etc.)
|
|
5
|
+
* so the server starts in <2s instead of >90s.
|
|
6
|
+
*
|
|
7
|
+
* Usage: node dist/mcp-entry.js [--vault <path>]
|
|
8
|
+
*/
|
|
9
|
+
import { resolve } from 'node:path';
|
|
10
|
+
import { startMcpServer } from './mcp-server.js';
|
|
11
|
+
const args = process.argv.slice(2);
|
|
12
|
+
let vault = process.cwd();
|
|
13
|
+
for (let i = 0; i < args.length; i++) {
|
|
14
|
+
const next = args[i + 1];
|
|
15
|
+
if (args[i] === '--vault' && next) {
|
|
16
|
+
vault = resolve(next);
|
|
17
|
+
break;
|
|
18
|
+
}
|
|
19
|
+
}
|
|
20
|
+
startMcpServer(vault);
|
|
21
|
+
//# sourceMappingURL=mcp-entry.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"mcp-entry.js","sourceRoot":"","sources":["../src/mcp-entry.ts"],"names":[],"mappings":";AAEA;;;;;;GAMG;AAEH,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AAEjD,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;AACnC,IAAI,KAAK,GAAG,OAAO,CAAC,GAAG,EAAE,CAAC;AAE1B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;IACrC,MAAM,IAAI,GAAG,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IACzB,IAAI,IAAI,CAAC,CAAC,CAAC,KAAK,SAAS,IAAI,IAAI,EAAE,CAAC;QAClC,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;QACtB,MAAM;IACR,CAAC;AACH,CAAC;AAED,cAAc,CAAC,KAAK,CAAC,CAAC"}
|
|
@@ -0,0 +1,15 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* WikiMem MCP Server — JSON-RPC 2.0 over stdin/stdout.
|
|
3
|
+
*
|
|
4
|
+
* Implements the Model Context Protocol without the SDK dependency.
|
|
5
|
+
* Reads newline-delimited JSON from stdin, writes responses to stdout.
|
|
6
|
+
*
|
|
7
|
+
* Tools exposed:
|
|
8
|
+
* wikimem_search — BM25 keyword search over wiki pages
|
|
9
|
+
* wikimem_read — Read a specific wiki page by title or filename
|
|
10
|
+
* wikimem_list — List all wiki pages with metadata
|
|
11
|
+
* wikimem_status — Vault statistics (page count, words, orphans, …)
|
|
12
|
+
* wikimem_ingest — Ingest a file or URL into the vault (async)
|
|
13
|
+
*/
|
|
14
|
+
export declare function startMcpServer(vault: string): Promise<void>;
|
|
15
|
+
//# sourceMappingURL=mcp-server.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"mcp-server.d.ts","sourceRoot":"","sources":["../src/mcp-server.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AA4YH,wBAAsB,cAAc,CAAC,KAAK,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,CA0CjE"}
|