Replace Ollama docx import with mammoth HTML conversion
Remove all AI processing from the import-docx endpoint. Instead: - Convert the .docx to HTML with mammoth.convertToHtml (preserves bold, headings, lists, etc.) — the result goes straight into the item body and renders in the lesson viewer - Extract raw text in parallel to pick the document title from the first non-empty line - No external services, no AI, works on the deployed server Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
nmemmertandClaude Sonnet 4.6 committed
1 parent
4100107c0a
commit
19f922c4ba
2 files changed
+25
-126
No files matched your search
@@ -1518,127 +1518,31 @@ app.get('/api/admin/homeschool-files', requirePin, (req, res) => {
|
||||
res.json({ files: files.map(({ name, path }) => ({ name, path })) });
|
||||
});
|
||||
|
||||
async function callOllamaTextLocal(fullPrompt) {
|
||||
const host = (getSetting('ollama_host') || 'http://localhost:11434').replace(/\/+$/, '');
|
||||
const model = getSetting('ollama_model') || 'llava';
|
||||
const controller = new AbortController();
|
||||
const timer = setTimeout(() => controller.abort(), 180_000);
|
||||
let r;
|
||||
try {
|
||||
r = await fetch(`${host}/api/generate`, {
|
||||
method: 'POST',
|
||||
headers: { 'Content-Type': 'application/json' },
|
||||
body: JSON.stringify({ model, prompt: fullPrompt, stream: false, format: 'json' }),
|
||||
signal: controller.signal,
|
||||
});
|
||||
} catch (err) {
|
||||
if (err.name === 'AbortError') throw new Error('Ollama timed out (3 min).');
|
||||
throw new Error(`Couldn't reach Ollama at ${host} — is it running? (${err.message})`);
|
||||
} finally { clearTimeout(timer); }
|
||||
if (!r.ok) {
|
||||
const t = await r.text().catch(() => '');
|
||||
if (r.status === 404) throw new Error(`Model "${model}" not found on Ollama. Run: ollama pull ${model}`);
|
||||
throw new Error(`Ollama error (${r.status}): ${t.slice(0, 200)}`);
|
||||
}
|
||||
const data = await r.json();
|
||||
return parseVisionJson(data.response);
|
||||
}
|
||||
|
||||
async function callOllamaTextCloud(fullPrompt) {
|
||||
const apiKey = getSetting('ollama_cloud_api_key') || '';
|
||||
if (!apiKey) throw new Error('Ollama Cloud API key not set — add it in Settings.');
|
||||
const host = (getSetting('ollama_host') || 'http://localhost:11434').replace(/\/+$/, '');
|
||||
const baseModel = (getSetting('ollama_cloud_model') || 'gemma4').replace(/:cloud$/, '');
|
||||
const model = `${baseModel}:cloud`;
|
||||
const controller = new AbortController();
|
||||
const timer = setTimeout(() => controller.abort(), 180_000);
|
||||
let r;
|
||||
try {
|
||||
r = await fetch(`${host}/api/chat`, {
|
||||
method: 'POST',
|
||||
headers: { 'Content-Type': 'application/json' },
|
||||
body: JSON.stringify({ model, messages: [{ role: 'user', content: fullPrompt }], stream: false }),
|
||||
signal: controller.signal,
|
||||
});
|
||||
} catch (err) {
|
||||
if (err.name === 'AbortError') throw new Error('Ollama Cloud timed out (3 min).');
|
||||
throw new Error(`Couldn't reach local Ollama at ${host} — is it running? (${err.message})`);
|
||||
} finally { clearTimeout(timer); }
|
||||
if (!r.ok) {
|
||||
const t = await r.text().catch(() => '');
|
||||
if (r.status === 404) throw new Error(`Cloud model "${model}" not found. Run: ollama pull ${model}`);
|
||||
throw new Error(`Ollama Cloud error (${r.status}): ${t.slice(0, 300)}`);
|
||||
}
|
||||
const data = await r.json();
|
||||
const content = data.message?.content;
|
||||
if (!content) throw new Error('Ollama Cloud returned an empty response.');
|
||||
return parseVisionJson(content);
|
||||
}
|
||||
|
||||
const TEXT_SCAN_PROMPTS = {
|
||||
lesson: `You are given the text content of a homeschool lesson or assignment document. Extract it as structured content.
|
||||
|
||||
Respond with ONLY a JSON object in exactly this shape, nothing else, no markdown fences:
|
||||
{"title": "the document title or heading", "body": "the full text content, paragraphs separated by a blank line"}`,
|
||||
|
||||
quiz: `You are given the text content of a homeschool quiz or test document. Extract every question exactly as written.
|
||||
|
||||
For each question determine:
|
||||
- "type": "mc" if it has lettered/numbered answer choices, "tf" if it is true/false, "short" if it expects a written answer
|
||||
- "prompt": the exact question text
|
||||
- "choices": for "mc", an array of the exact answer choice texts (empty array for "tf" or "short")
|
||||
- "correctAnswer": the marked answer if clearly indicated; for "tf" use exactly "true" or "false"; otherwise use ""
|
||||
- "points": 1 unless the document states a different value
|
||||
|
||||
Respond with ONLY a JSON object in exactly this shape, nothing else, no markdown fences:
|
||||
{"title": "the quiz/test title", "questions": [{"type": "mc", "prompt": "...", "choices": ["...", "..."], "correctAnswer": "...", "points": 1}]}`,
|
||||
};
|
||||
|
||||
app.post('/api/admin/import-docx', requirePin, async (req, res) => {
|
||||
const { path: relPath, base64: b64, mode } = req.body;
|
||||
if (!TEXT_SCAN_PROMPTS[mode] || (!relPath && !b64)) {
|
||||
return res.status(400).json({ error: 'mode and either path or base64 required' });
|
||||
}
|
||||
const { base64: b64 } = req.body;
|
||||
if (!b64) return res.status(400).json({ error: 'base64 required' });
|
||||
|
||||
let buffer;
|
||||
if (b64) {
|
||||
buffer = Buffer.from(b64, 'base64');
|
||||
} else {
|
||||
const fullPath = join(HOMESCHOOL_DIR, relPath);
|
||||
if (!fullPath.startsWith(HOMESCHOOL_DIR)) {
|
||||
return res.status(403).json({ error: 'Access denied' });
|
||||
}
|
||||
try { buffer = await readFile(fullPath); }
|
||||
catch (err) { return res.status(404).json({ error: `Could not read file: ${err.message}` }); }
|
||||
}
|
||||
try { buffer = Buffer.from(b64, 'base64'); }
|
||||
catch (err) { return res.status(400).json({ error: 'Invalid base64' }); }
|
||||
|
||||
let docText;
|
||||
let html, rawText;
|
||||
try {
|
||||
const result = await mammoth.extractRawText({ buffer });
|
||||
docText = result.value.trim();
|
||||
const [htmlResult, textResult] = await Promise.all([
|
||||
mammoth.convertToHtml({ buffer }),
|
||||
mammoth.extractRawText({ buffer }),
|
||||
]);
|
||||
html = htmlResult.value.trim();
|
||||
rawText = textResult.value.trim();
|
||||
} catch (err) {
|
||||
return res.status(500).json({ error: `Could not parse docx: ${err.message}` });
|
||||
}
|
||||
if (!docText) return res.status(400).json({ error: 'Document appears to be empty' });
|
||||
if (!rawText) return res.status(400).json({ error: 'Document appears to be empty' });
|
||||
|
||||
const fullPrompt = TEXT_SCAN_PROMPTS[mode] + '\n\nDOCUMENT TEXT:\n' + docText;
|
||||
|
||||
const primary = getSetting('scan_primary') || 'local';
|
||||
const providerFns = { local: callOllamaTextLocal, ollama_cloud: callOllamaTextCloud };
|
||||
const order = primary === 'ollama_cloud' ? ['ollama_cloud', 'local'] : ['local', 'ollama_cloud'];
|
||||
const isConfigured = { local: true, ollama_cloud: !!getSetting('ollama_cloud_api_key') };
|
||||
|
||||
const errors = [];
|
||||
for (const provider of order) {
|
||||
if (!isConfigured[provider]) continue;
|
||||
try {
|
||||
const result = await providerFns[provider](fullPrompt);
|
||||
return res.json(result);
|
||||
} catch (err) {
|
||||
errors.push(`${provider}: ${err.message}`);
|
||||
}
|
||||
}
|
||||
return res.status(502).json({ error: errors.join(' | ') || 'No AI providers configured.' });
|
||||
// Use first non-empty line of raw text as the title
|
||||
const firstLine = rawText.split('\n').map(l => l.trim()).find(Boolean) || '';
|
||||
return res.json({ title: firstLine, body: html });
|
||||
});
|
||||
|
||||
app.get('/health', (req, res) => res.json({ ok: true }));
|
||||
|
||||
Reference in new issue
Block a user