storme/nexusAI
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

utility inference layer 3

Browse Source
This commit is contained in:
Storme-bit
2026-08-17 10:11:08 -07:00
parent b9195921dc
commit 67e2b446e4
3 changed files with 209 additions and 87 deletions
Download Patch File Download Diff File Expand all files Collapse all files
packages/memory-service/src/summarization/project.js
+10 -54
View File
@@ -1,4 +1,4 @@
const { SERVICES, getEnv, SUMMARIES } = require('@nexusai/shared');
const { SERVICES, getEnv, SUMMARIES, utilityInference } = require('@nexusai/shared');
const {
getSessionSummariesForProject,
getProjectOverviewSummary,
@@ -9,9 +9,6 @@ const {
const { getEpisodesByProject } = require('../episodic');
const { getProject } = require('../db/projects');
const EXTRACTION_URL = getEnv('EXTRACTION_URL', 'http://localhost:11434');
const EXTRACTION_MODEL = getEnv('EXTRACTION_MODEL', 'qwen2.5:3b');
const MAX_SUMMARY_CHARS = SUMMARIES.MAX_SUMMARY_CHARS; // generous ceiling before we truncate input
function buildProjectSummaryPrompt(projectName, sessionSummaries) {
@@ -24,8 +21,9 @@ function buildProjectSummaryPrompt(projectName, sessionSummaries) {
summaryBlock = summaryBlock.slice(-MAX_SUMMARY_CHARS);
}
// No ChatML wrapper — the model's own prompt template is applied server-side
// by the inference service's /utility/complete route (Ollama /api/chat).
return [
'<|im_start|>user',
`The following are session summaries from a project called "${projectName}".`,
'Write a project overview covering: goals, progress, key decisions, and current state.',
'Scale the length to the material — use multiple paragraphs for complex projects, a few sentences for simple ones.',
@@ -33,8 +31,6 @@ function buildProjectSummaryPrompt(projectName, sessionSummaries) {
'Write in third person. Output only the overview text, no headings or labels.',
'',
summaryBlock,
'<|im_end|>',
'<|im_start|>assistant',
].join('\n');
}
@@ -49,8 +45,9 @@ function buildProjectSummaryFromEpisodesPrompt(projectName, episodes) {
episodeBlock = episodeBlock.slice(-MAX_SUMMARY_CHARS);
}
// No ChatML wrapper — the model's own prompt template is applied server-side
// by the inference service's /utility/complete route (Ollama /api/chat).
return [
'<|im_start|>user',
`The following are conversations from a project called "${projectName}".`,
'Write a project overview covering: goals, progress, key decisions, and current state.',
'Scale the length to the material — use multiple paragraphs for complex projects, a few sentences for simple ones.',
@@ -58,58 +55,17 @@ function buildProjectSummaryFromEpisodesPrompt(projectName, episodes) {
'Write in third person. Output only the overview text, no headings or labels.',
'',
episodeBlock,
'<|im_end|>',
'<|im_start|>assistant',
].join('\n');
}
async function generateProjectSummaryFromEpisodes(projectName, episodes) {
const prompt = buildProjectSummaryFromEpisodesPrompt(projectName, episodes);
const res = await fetch(`${EXTRACTION_URL}/api/generate`, {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({
model: EXTRACTION_MODEL,
prompt,
stream: false,
options: { temperature: 0.2, num_predict: 1200 },
}),
});
if (!res.ok) throw new Error(`Ollama responded ${res.status}`);
const data = await res.json();
const raw = data.response?.trim() ?? '';
return raw
.replace(/<\|im_start\|>.*?<\|im_end\|>/gs, '')
.replace(/<\|im_start\|>|<\|im_end\|>|<\|im_sep\|>/g, '')
.trim();
const user = buildProjectSummaryFromEpisodesPrompt(projectName, episodes);
return utilityInference({ user, temperature: 0.2, maxTokens: 1200 });
}
async function generateProjectSummary(projectName, sessionSummaries) {
const prompt = buildProjectSummaryPrompt(projectName, sessionSummaries);
const res = await fetch(`${EXTRACTION_URL}/api/generate`, {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({
model: EXTRACTION_MODEL,
prompt,
stream: false,
// No format: 'json' — we want free-text narrative, same as session summarization
options: { temperature: 0.2, num_predict: 1200 },
}),
});
if (!res.ok) throw new Error(`Ollama responded ${res.status}`);
const data = await res.json();
const raw = data.response?.trim() ?? '';
return raw
.replace(/<\|im_start\|>.*?<\|im_end\|>/gs, '')
.replace(/<\|im_start\|>|<\|im_end\|>|<\|im_sep\|>/g, '')
.trim();
const user = buildProjectSummaryPrompt(projectName, sessionSummaries);
return utilityInference({ user, temperature: 0.2, maxTokens: 1200 });
}
// Main entry point — called by the route handler
@@ -142,4 +98,4 @@ async function generateAndStoreProjectSummary(projectId) {
}
}
module.exports = { generateAndStoreProjectSummary };
module.exports = { generateAndStoreProjectSummary };