@maintainabilityai/research-runner 0.1.14 → 0.1.15
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.
|
@@ -40,7 +40,10 @@ async function runHackerNewsSearch(opts) {
|
|
|
40
40
|
fromQuery: query,
|
|
41
41
|
title: r.title,
|
|
42
42
|
url,
|
|
43
|
-
|
|
43
|
+
// Show HN / Ask HN posts carry a self-post body. URL-only
|
|
44
|
+
// submissions have no body — leave empty; the synth agent
|
|
45
|
+
// can read the linked URL if it needs more context.
|
|
46
|
+
content: r.storyText,
|
|
44
47
|
score: pointsToScore(r.points),
|
|
45
48
|
publishedDate: r.createdAt || undefined,
|
|
46
49
|
authors: r.author ? [r.author] : undefined,
|
|
@@ -15,6 +15,8 @@ export interface HackerNewsResult {
|
|
|
15
15
|
points: number;
|
|
16
16
|
numComments: number;
|
|
17
17
|
createdAt: string;
|
|
18
|
+
/** Self-post body for Show HN / Ask HN; empty for URL submissions. */
|
|
19
|
+
storyText: string;
|
|
18
20
|
}
|
|
19
21
|
export interface HackerNewsSearchOpts {
|
|
20
22
|
query: string;
|
|
@@ -39,6 +39,29 @@ async function hackerNewsSearch(opts) {
|
|
|
39
39
|
points: h.points ?? 0,
|
|
40
40
|
numComments: h.num_comments ?? 0,
|
|
41
41
|
createdAt: h.created_at ?? '',
|
|
42
|
+
// Algolia returns the self-post body for Show HN / Ask HN. URL-only
|
|
43
|
+
// submissions have this empty. Strip basic HTML tags so the excerpt
|
|
44
|
+
// is readable in the issue comment.
|
|
45
|
+
storyText: stripBasicHtml(h.story_text ?? '').slice(0, 2000),
|
|
42
46
|
})).filter(r => r.objectId && r.title);
|
|
43
47
|
return { query: opts.query, results, responseBytes: Buffer.byteLength(rawText, 'utf8'), httpStatus };
|
|
44
48
|
}
|
|
49
|
+
/**
|
|
50
|
+
* HN Algolia returns Show HN / Ask HN bodies with light HTML
|
|
51
|
+
* (`<p>`, `<i>`, etc.). Strip tags + decode the common entities so
|
|
52
|
+
* the excerpt blockquote in the issue body reads cleanly. No need for
|
|
53
|
+
* a full HTML parser — these posts are plain text with a sprinkle of
|
|
54
|
+
* inline tags.
|
|
55
|
+
*/
|
|
56
|
+
function stripBasicHtml(s) {
|
|
57
|
+
return s
|
|
58
|
+
.replace(/<\/?(?:p|br|i|b|em|strong|code|pre|a|ul|ol|li)[^>]*>/gi, ' ')
|
|
59
|
+
.replace(/ /g, ' ')
|
|
60
|
+
.replace(/&/g, '&')
|
|
61
|
+
.replace(/</g, '<')
|
|
62
|
+
.replace(/>/g, '>')
|
|
63
|
+
.replace(/"/g, '"')
|
|
64
|
+
.replace(/'/g, "'")
|
|
65
|
+
.replace(/\s+/g, ' ')
|
|
66
|
+
.trim();
|
|
67
|
+
}
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@maintainabilityai/research-runner",
|
|
3
|
-
"version": "0.1.
|
|
3
|
+
"version": "0.1.15",
|
|
4
4
|
"description": "Research + PRD agent runner — orchestrates the Archeologist and PRD pipelines for the MaintainabilityAI governance mesh",
|
|
5
5
|
"license": "MIT",
|
|
6
6
|
"author": "MaintainabilityAI",
|