indexx.dev / tweets2bsky
forked from j4ck.xyz/tweets2bsky
A simple tool which lets you scrape twitter accounts and crosspost them to bluesky accounts! Comes with a CLI and a webapp for managing profiles! Works with images/videos/link embeds/threads.
fork atom

feat: link non-bsky handles to twitter

Jack G 86015a55 60a3c3e3

+53
unified split
+53
src/index.ts
··· 1075 1075 return chunks; 1076 1076 } 1077 1077 1078 + function utf16IndexToUtf8Index(text: string, index: number): number { 1079 + return Buffer.byteLength(text.slice(0, index), 'utf8'); 1080 + } 1081 + 1082 + function rangesOverlap(startA: number, endA: number, startB: number, endB: number): boolean { 1083 + return startA < endB && startB < endA; 1084 + } 1085 + 1086 + function addTwitterHandleLinkFacets(text: string, facets?: any[]): any[] | undefined { 1087 + const existingFacets = facets ?? []; 1088 + const newFacets: any[] = []; 1089 + const regex = /@([A-Za-z0-9_]{1,15})/g; 1090 + let match: RegExpExecArray | null; 1091 + 1092 + while ((match = regex.exec(text))) { 1093 + const handle = match[1]; 1094 + if (!handle) continue; 1095 + 1096 + const atIndex = match.index; 1097 + const prevChar = atIndex > 0 ? text[atIndex - 1] : ''; 1098 + if (prevChar && /[A-Za-z0-9_]/.test(prevChar)) continue; 1099 + 1100 + const endIndex = atIndex + handle.length + 1; 1101 + const trailing = text.slice(endIndex); 1102 + if (trailing.startsWith('.') && /^\.[A-Za-z0-9-]+/.test(trailing)) continue; 1103 + 1104 + const nextChar = endIndex < text.length ? text[endIndex] : ''; 1105 + if (nextChar && /[A-Za-z0-9_]/.test(nextChar)) continue; 1106 + 1107 + const byteStart = utf16IndexToUtf8Index(text, atIndex); 1108 + const byteEnd = utf16IndexToUtf8Index(text, endIndex); 1109 + 1110 + const overlaps = existingFacets.some((facet) => 1111 + rangesOverlap(byteStart, byteEnd, facet.index.byteStart, facet.index.byteEnd), 1112 + ); 1113 + if (overlaps) continue; 1114 + 1115 + newFacets.push({ 1116 + index: { byteStart, byteEnd }, 1117 + features: [ 1118 + { 1119 + $type: 'app.bsky.richtext.facet#link', 1120 + uri: `https://twitter.com/${handle}`, 1121 + }, 1122 + ], 1123 + }); 1124 + } 1125 + 1126 + if (newFacets.length === 0) return facets; 1127 + return [...existingFacets, ...newFacets].sort((a, b) => a.index.byteStart - b.index.byteStart); 1128 + } 1129 + 1078 1130 // Simple p-limit implementation for concurrency control 1079 1131 const pLimit = (concurrency: number) => { 1080 1132 const queue: (() => Promise<void>)[] = []; ··· 1670 1722 1671 1723 const rt = new RichText({ text: chunk }); 1672 1724 await rt.detectFacets(agent); 1725 + rt.facets = addTwitterHandleLinkFacets(rt.text, rt.facets); 1673 1726 const detectedLangs = detectLanguage(chunk); 1674 1727 1675 1728 // Add offset for split chunks to ensure correct ordering/threading