Phase 3: PDF/DOCX extraction, chunking, LLM client with mock interface

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

internal/parse/acceptance_test.go

@@ -0,0 +1,53 @@
+package parse_test
+
+import (
+	"bytes"
+	"strings"
+	"testing"
+
+	"qbank/internal/parse"
+)
+
+// TestAcceptanceDOCXPipeline verifies the full DOCX → text → chunk pipeline
+// using a handcrafted in-memory docx with known content.
+func TestAcceptanceDOCXPipeline(t *testing.T) {
+	const docXML = `<?xml version="1.0" encoding="UTF-8"?>
+<w:document xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main">
+  <w:body>
+    <w:p><w:r><w:t>1. Which keyword declares a variable in Go?</w:t></w:r></w:p>
+    <w:p><w:r><w:t>A) var</w:t></w:r></w:p>
+    <w:p><w:r><w:t>B) let</w:t></w:r></w:p>
+    <w:p><w:r><w:t>C) dim</w:t></w:r></w:p>
+    <w:p><w:r><w:t>Correct: A</w:t></w:r></w:p>
+    <w:p><w:r><w:t>2. What does fmt.Println return?</w:t></w:r></w:p>
+    <w:p><w:r><w:t>A) Nothing</w:t></w:r></w:p>
+    <w:p><w:r><w:t>B) n int, err error</w:t></w:r></w:p>
+    <w:p><w:r><w:t>Correct: B</w:t></w:r></w:p>
+  </w:body>
+</w:document>`
+
+	docx := buildDocx(t, docXML)
+
+	text, err := parse.ExtractDOCX(bytes.NewReader(docx))
+	if err != nil {
+		t.Fatalf("ExtractDOCX: %v", err)
+	}
+
+	wantPhrases := []string{
+		"Which keyword declares a variable",
+		"fmt.Println",
+		"n int, err error",
+	}
+	for _, phrase := range wantPhrases {
+		if !strings.Contains(text, phrase) {
+			t.Errorf("text missing %q\nfull text:\n%s", phrase, text)
+		}
+	}
+
+	// Chunking should produce at least 1 chunk.
+	chunks := parse.Chunk(text, 10_000)
+	if len(chunks) == 0 {
+		t.Error("Chunk returned 0 chunks for non-empty text")
+	}
+	t.Logf("extracted %d chars, %d chunk(s)", len(text), len(chunks))
+}