@lota-sdk/core 0.4.8 → 0.4.10

package/src/storage/attachment-parser.ts

@@ -1,9 +1,25 @@
 import { SUPPORTED_ATTACHMENT_MIME_TYPES, getAttachmentExtension } from '@lota-sdk/shared'
+import { Schema, Effect } from 'effect'
 import mammoth from 'mammoth'
 import { PDFParse } from 'pdf-parse'
 
 const READ_FILE_PARTS_CHARS_PER_PAGE = 3500
 
+class AttachmentParserError extends Schema.TaggedErrorClass<AttachmentParserError>()('AttachmentParserError', {
+  message: Schema.String,
+  cause: Schema.Defect,
+}) {}
+
+function tryAttachmentPromise<A>(
+  message: string,
+  thunk: () => PromiseLike<A>,
+): Effect.Effect<A, AttachmentParserError> {
+  return Effect.tryPromise({
+    try: () => Promise.resolve(thunk()),
+    catch: (cause) => new AttachmentParserError({ message, cause }),
+  })
+}
+
 export function normalizeExtractedText(value: string): string {
   return value.replaceAll('\u0000', '').replace(/\r\n/g, '\n').trim()
 }
@@ -29,44 +45,70 @@ export function isTextAttachmentFile(file: File): boolean {
   return ext === 'txt' || ext === 'md' || ext === 'markdown'
 }
 
-export async function extractPdfPages(file: File): Promise<string[]> {
-  const parser = new PDFParse({ data: new Uint8Array(await file.arrayBuffer()) })
-  try {
-    const info = await parser.getInfo()
-    const textResult = await parser.getText()
-
-    const pageTextByNumber = new Map<number, string>()
-    for (const page of textResult.pages) {
-      pageTextByNumber.set(page.num, normalizeExtractedText(page.text))
-    }
-
-    const pages: string[] = []
-    for (let pageNumber = 1; pageNumber <= info.total; pageNumber += 1) {
-      pages.push(pageTextByNumber.get(pageNumber) ?? '')
-    }
-    return pages
-  } finally {
-    await parser.destroy()
-  }
+export function extractPdfPages(file: File): Promise<string[]> {
+  return Effect.runPromise(
+    Effect.gen(function* () {
+      const buffer = yield* tryAttachmentPromise('Failed to read PDF attachment buffer.', () => file.arrayBuffer())
+      const parser = new PDFParse({ data: new Uint8Array(buffer) })
+
+      return yield* Effect.gen(function* () {
+        const info = yield* tryAttachmentPromise('Failed to read PDF attachment metadata.', () => parser.getInfo())
+        const textResult = yield* tryAttachmentPromise('Failed to extract PDF attachment text.', () => parser.getText())
+        const pageTextByNumber = new Map<number, string>()
+        for (const page of textResult.pages) {
+          pageTextByNumber.set(page.num, normalizeExtractedText(page.text))
+        }
+
+        const pages: string[] = []
+        for (let pageNumber = 1; pageNumber <= info.total; pageNumber += 1) {
+          pages.push(pageTextByNumber.get(pageNumber) ?? '')
+        }
+        return pages
+      }).pipe(
+        Effect.ensuring(
+          tryAttachmentPromise('Failed to destroy PDF parser.', () => parser.destroy()).pipe(
+            Effect.orDie,
+            Effect.asVoid,
+          ),
+        ),
+      )
+    }),
+  )
 }
 
-export async function extractDocxText(file: File): Promise<string> {
-  const buffer = Buffer.from(await file.arrayBuffer())
-  const result = await mammoth.extractRawText({ buffer })
-  return normalizeExtractedText(result.value)
+export function extractDocxText(file: File): Promise<string> {
+  return Effect.runPromise(
+    Effect.gen(function* () {
+      const arrayBuffer = yield* tryAttachmentPromise('Failed to read DOCX attachment buffer.', () =>
+        file.arrayBuffer(),
+      )
+      const result = yield* tryAttachmentPromise('Failed to extract DOCX attachment text.', () =>
+        mammoth.extractRawText({ buffer: Buffer.from(arrayBuffer) }),
+      )
+      return normalizeExtractedText(result.value)
+    }),
+  )
 }
 
-export async function extractAttachmentText(file: File): Promise<string> {
+export function extractAttachmentText(file: File): Promise<string> {
   if (isTextAttachmentFile(file)) {
-    return normalizeExtractedText(await file.text())
+    return Effect.runPromise(
+      tryAttachmentPromise('Failed to read text attachment.', () => file.text()).pipe(
+        Effect.map(normalizeExtractedText),
+      ),
+    )
   }
   if (isPdfAttachmentFile(file)) {
-    return normalizeExtractedText((await extractPdfPages(file)).join('\n\n'))
+    return Effect.runPromise(
+      tryAttachmentPromise('Failed to extract PDF attachment text.', () => extractPdfPages(file)).pipe(
+        Effect.map((pages) => normalizeExtractedText(pages.join('\n\n'))),
+      ),
+    )
   }
   if (isDocxAttachmentFile(file)) {
     return extractDocxText(file)
   }
-  return ''
+  return Promise.resolve('')
 }
 
 export function splitExtractedTextIntoPages(extractedText: string): string[] {