feat: port contexts layer and MCP parity
This commit is contained in:
“BeeRad”
@@ -1,12 +1,13 @@
|
||||
import { openai as openaiProvider } from '@ai-sdk/openai';
|
||||
import { generateText } from 'ai';
|
||||
import { hasValidOpenAiKey } from '../storage/apiKeys';
|
||||
import type { CanonicalNodeMetadata } from '@/types/database';
|
||||
|
||||
export interface DescriptionInput {
|
||||
title: string;
|
||||
source?: string;
|
||||
link?: string;
|
||||
metadata?: {
|
||||
metadata?: CanonicalNodeMetadata & {
|
||||
source?: string;
|
||||
channel_name?: string;
|
||||
author?: string;
|
||||
@@ -18,52 +19,12 @@ export interface DescriptionInput {
|
||||
dimensions?: string[];
|
||||
}
|
||||
|
||||
// Re-export for backwards compatibility — canonical source is ../storage/apiKeys
|
||||
export { hasValidOpenAiKey } from '../storage/apiKeys';
|
||||
|
||||
/**
|
||||
* Generate a simple fallback description without AI.
|
||||
* Used when no API key is available or for simple inputs.
|
||||
*/
|
||||
export function generateFallbackDescription(input: DescriptionInput): string {
|
||||
const { title, metadata, dimensions } = input;
|
||||
|
||||
// Build a contextual fallback
|
||||
const parts: string[] = [];
|
||||
|
||||
if (metadata?.author || metadata?.channel_name) {
|
||||
parts.push(`By ${metadata.author || metadata.channel_name}`);
|
||||
}
|
||||
|
||||
if (dimensions?.length) {
|
||||
parts.push(`in ${dimensions.slice(0, 2).join(', ')}`);
|
||||
}
|
||||
|
||||
if (parts.length > 0) {
|
||||
return `${parts.join(' — ')}: ${title.slice(0, 200)}`;
|
||||
}
|
||||
|
||||
return title.slice(0, 280);
|
||||
}
|
||||
|
||||
/**
|
||||
* Generate a 280-character description for a knowledge node.
|
||||
* Contextually grounded - adapts to node type (person, concept, article, etc.)
|
||||
*
|
||||
* IMPORTANT: Returns fallback immediately if no valid API key is configured.
|
||||
* This prevents slow node creation (9-13s timeout) when OpenAI is unavailable.
|
||||
*/
|
||||
export async function generateDescription(input: DescriptionInput): Promise<string> {
|
||||
// Fast path: skip AI if no valid API key
|
||||
if (!hasValidOpenAiKey()) {
|
||||
console.log(`[DescriptionService] No valid OpenAI key, using fallback for: "${input.title}"`);
|
||||
return generateFallbackDescription(input);
|
||||
}
|
||||
|
||||
// Fast path: skip AI for very short inputs (likely just notes)
|
||||
if (!input.source && !input.link && input.title.length < 30) {
|
||||
console.log(`[DescriptionService] Short input, using fallback for: "${input.title}"`);
|
||||
return generateFallbackDescription(input);
|
||||
return `${input.title}. Added via Quick Add with no further context yet, so the reason it belongs in the graph is not fully inferred. It has not been reviewed yet.`.slice(0, 500);
|
||||
}
|
||||
|
||||
try {
|
||||
@@ -78,31 +39,38 @@ export async function generateDescription(input: DescriptionInput): Promise<stri
|
||||
temperature: 0.3,
|
||||
});
|
||||
|
||||
const finalDescription = sanitizeDescription(response.text, input);
|
||||
const description = sanitizeDescription(response.text, input);
|
||||
|
||||
console.log(`[DescriptionService] Generated: "${finalDescription}"`);
|
||||
console.log(`[DescriptionService] Generated: "${description}"`);
|
||||
|
||||
return finalDescription;
|
||||
return description;
|
||||
} catch (error) {
|
||||
console.error('[DescriptionService] Error generating description:', error);
|
||||
// Return a fallback description
|
||||
return generateFallbackDescription(input);
|
||||
return `${input.title}. Added via Quick Add with no further context yet, so the reason it belongs in the graph is not fully inferred. It has not been reviewed yet.`.slice(0, 500);
|
||||
}
|
||||
}
|
||||
|
||||
function buildDescriptionPrompt(input: DescriptionInput): string {
|
||||
const normalizedSource = (input.metadata?.source || '').toLowerCase();
|
||||
const sourceMetadata = input.metadata?.source_metadata as Record<string, unknown> | undefined;
|
||||
const sourceType = typeof input.metadata?.type === 'string'
|
||||
? input.metadata.type
|
||||
: typeof input.metadata?.source === 'string'
|
||||
? input.metadata.source
|
||||
: '';
|
||||
const normalizedSource = sourceType.toLowerCase();
|
||||
const url = typeof input.link === 'string' ? input.link.trim() : '';
|
||||
|
||||
// Best-effort creator hint from structured metadata (when available),
|
||||
// but never assume a particular extraction source (YouTube vs paper vs website vs note).
|
||||
const creatorHint =
|
||||
input.metadata?.author?.trim() ||
|
||||
input.metadata?.channel_name?.trim() ||
|
||||
(typeof sourceMetadata?.author === 'string' ? sourceMetadata.author.trim() : '') ||
|
||||
(typeof sourceMetadata?.channel_name === 'string' ? sourceMetadata.channel_name.trim() : '') ||
|
||||
(typeof input.metadata?.author === 'string' ? input.metadata.author.trim() : '') ||
|
||||
(typeof input.metadata?.channel_name === 'string' ? input.metadata.channel_name.trim() : '') ||
|
||||
'';
|
||||
|
||||
// Best-effort publisher / container hint (less ideal than a true author, but better than nothing).
|
||||
const publisherHint = input.metadata?.site_name?.trim() || '';
|
||||
const publisherHint =
|
||||
(typeof sourceMetadata?.site_name === 'string' ? sourceMetadata.site_name.trim() : '') ||
|
||||
(typeof input.metadata?.site_name === 'string' ? input.metadata.site_name.trim() : '') ||
|
||||
'';
|
||||
|
||||
const likelyExternal =
|
||||
Boolean(url) ||
|
||||
@@ -123,38 +91,41 @@ function buildDescriptionPrompt(input: DescriptionInput): string {
|
||||
|
||||
if (input.link) lines.push(`URL: ${input.link}`);
|
||||
if (input.dimensions?.length) lines.push(`Dimensions: ${input.dimensions.join(', ')}`);
|
||||
if (input.metadata?.channel_name) lines.push(`Channel: ${input.metadata.channel_name}`);
|
||||
if (input.metadata?.author) lines.push(`Author: ${input.metadata.author}`);
|
||||
if (input.metadata?.site_name) lines.push(`Site: ${input.metadata.site_name}`);
|
||||
if (input.metadata?.source) lines.push(`Source type: ${input.metadata.source}`);
|
||||
if (input.metadata?.original_filename) lines.push(`Original filename: ${input.metadata.original_filename}`);
|
||||
if (typeof input.metadata?.pages === 'number') lines.push(`Pages: ${input.metadata.pages}`);
|
||||
if (typeof input.metadata?.text_length === 'number') lines.push(`Text length: ${input.metadata.text_length}`);
|
||||
if (sourceMetadata?.channel_name || input.metadata?.channel_name) lines.push(`Channel: ${sourceMetadata?.channel_name || input.metadata?.channel_name}`);
|
||||
if (sourceMetadata?.author || input.metadata?.author) lines.push(`Author: ${sourceMetadata?.author || input.metadata?.author}`);
|
||||
if (sourceMetadata?.site_name || input.metadata?.site_name) lines.push(`Site: ${sourceMetadata?.site_name || input.metadata?.site_name}`);
|
||||
if (sourceType) lines.push(`Source type: ${sourceType}`);
|
||||
if (sourceMetadata?.original_filename || input.metadata?.original_filename) lines.push(`Original filename: ${sourceMetadata?.original_filename || input.metadata?.original_filename}`);
|
||||
if (typeof sourceMetadata?.pages === 'number' || typeof input.metadata?.pages === 'number') lines.push(`Pages: ${sourceMetadata?.pages || input.metadata?.pages}`);
|
||||
if (typeof sourceMetadata?.text_length === 'number' || typeof input.metadata?.text_length === 'number') lines.push(`Text length: ${sourceMetadata?.text_length || input.metadata?.text_length}`);
|
||||
if (creatorHint) lines.push(`Creator hint: ${creatorHint}`);
|
||||
if (publisherHint) lines.push(`Publisher hint: ${publisherHint}`);
|
||||
lines.push(`Likely user-authored: ${likelyUserAuthored ? 'yes' : 'no'}`);
|
||||
|
||||
const contentPreview = input.source?.slice(0, 800) || '';
|
||||
if (contentPreview) lines.push(`Source excerpt: ${contentPreview}${input.source && input.source.length > 800 ? '...' : ''}`);
|
||||
const sourcePreview = input.source?.slice(0, 800) || '';
|
||||
if (sourcePreview) lines.push(`Source excerpt: ${sourcePreview}${input.source && input.source.length > 800 ? '...' : ''}`);
|
||||
|
||||
return `Write a description for this knowledge node. Max 280 characters.
|
||||
return `Write a natural description for this knowledge node. Max 500 characters.
|
||||
|
||||
Say WHAT this literally is and WHY it matters. Be concrete and specific — like you're telling a friend what this thing is in one breath.
|
||||
The description should read like normal prose, not a template or checklist. In one compact paragraph or a few natural sentences, make sure it clearly conveys:
|
||||
1) what this literally is
|
||||
2) why it is in Brad's graph
|
||||
3) its current status in Brad's workflow
|
||||
|
||||
RULES:
|
||||
1) Name the format only if the context clearly supports it: "Podcast episode where…", "Blog post arguing…", "Personal note capturing…", "Research paper showing…", "Resume/CV for…", "Document likely containing…", "Idea that…"
|
||||
2) Name people by role — channel/host is the creator, title figures are guests/subjects. Use the Creator hint if available.
|
||||
3) State the actual claim, finding, or insight from the content — not a vague summary of the topic.
|
||||
4) End with why it's interesting or important — one concrete phrase.
|
||||
5) ABSOLUTELY FORBIDDEN — these words will be rejected: "discusses", "explores", "examines", "talks about", "is about", "delves into", "emphasizing the need for". State things directly instead.
|
||||
6) Do NOT start with "Your note —" or "This note —". Use a concrete opener tied to the actual artifact.
|
||||
7) If the artifact type is unclear, say so explicitly using words like "likely", "appears to be", or "unclear" rather than guessing a confident format.
|
||||
4) If the reason it belongs in the graph cannot be inferred from title, source excerpt, URL, metadata, or dimensions, say that naturally rather than inventing context.
|
||||
5) If workflow status is unknown, say that naturally, for example by noting it has not been reviewed yet.
|
||||
6) Do NOT use labels or headings like "WHAT:", "WHY:", or "STATUS:".
|
||||
7) ABSOLUTELY FORBIDDEN — these words and phrases will be rejected: "discusses", "explores", "examines", "talks about", "is about", "delves into", "emphasizing the need for", "insightful for understanding", "relevant to", "important for", "useful for understanding". State things directly instead.
|
||||
8) Do NOT start with "Your note —" or "This note —". Use a concrete opener tied to the actual artifact.
|
||||
9) If the artifact type is unclear, say so explicitly using words like "likely", "appears to be", or "unclear" rather than guessing a confident format.
|
||||
|
||||
GOOD: "Karpathy blog post — AI agents make software fluid, ripping functionality from repos instead of taking dependencies. Signals the end of monolithic libraries."
|
||||
GOOD: "Dwarkesh Patel interview with Anthropic CEO Dario Amodei — argues we're nearing the end of exponential AI scaling. Key signal for what comes next."
|
||||
GOOD: "Personal note capturing a recurring pattern: morning optimism reverses to evening pessimism. Indicates a belief-level swing worth tracking."
|
||||
GOOD: "Resume/CV for Brad Morris outlining work in AI systems, context engineering, and RA-H. Useful as a compact record of background, projects, and expertise."
|
||||
GOOD: "Document likely related to Brad Morris's work history and AI consulting, but the exact artifact type is unclear from the available context. Still useful as a reference profile."
|
||||
GOOD: "CS153 lecture by ElevenLabs co-founder Mati Staniszewski on production AI voice systems. Brad likely saved it as a follow-on to his interest in the ElevenLabs voice pipeline after CS153 ep.1, and it has not been reviewed yet."
|
||||
GOOD: "YouTube talk by Lex Fridman with Sam Altman on AGI timelines and OpenAI strategy. It was added via Quick Add and the exact reason it belongs in the graph is not yet inferred from the available context, and it has not been reviewed yet."
|
||||
GOOD: "Personal note capturing a recurring pattern: morning optimism reverses to evening pessimism. It belongs in the graph because it points to a belief-level pattern worth tracking against Brad's decision quality, and it has already been processed."
|
||||
BAD: "By Dario Amodei — discusses reaching the limits of exponential growth in AI, emphasizing the need for a critical perspective on future advancements."
|
||||
BAD: "This article explores ideas about how software is changing."
|
||||
|
||||
@@ -170,7 +141,7 @@ function sanitizeDescription(rawText: string, input: DescriptionInput): string {
|
||||
.replace(/^["']|["']$/g, '');
|
||||
|
||||
if (!singleLine) {
|
||||
return input.title.slice(0, 280);
|
||||
return `${input.title}. Added via Quick Add with no further context yet, so the reason it belongs in the graph is not fully inferred. It has not been reviewed yet.`.slice(0, 500);
|
||||
}
|
||||
|
||||
const noGenericPrefix = singleLine.replace(
|
||||
@@ -178,7 +149,7 @@ function sanitizeDescription(rawText: string, input: DescriptionInput): string {
|
||||
'Personal note capturing '
|
||||
);
|
||||
|
||||
return noGenericPrefix.slice(0, 280);
|
||||
return noGenericPrefix.slice(0, 500);
|
||||
}
|
||||
|
||||
export const descriptionService = {
|
||||
|
||||
Reference in New Issue
Block a user