Dev to main

.github/workflows/main.yml

@@ -1,57 +1,62 @@
 name: Backend CI
 
-
 on:
   push:
-    branches: ["main", "dev"]
+    branches: [main, dev]
   pull_request:
-    branches: ["main", "dev"]
-
+    branches: [main, dev]
 
 jobs:
   tests:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
         uses: actions/checkout@v6
+
       - name: Set up Python
         uses: actions/setup-python@v6
         with:
           python-version: "3.12"
+
       - name: Install dependencies
         run: |
           pip install --upgrade pip
           pip install --no-cache-dir -r requirements.txt
+
       - name: Lint with Pylint
         run: |
           pylint **/*.py
+
       - name: Unittests + coverage
         run: |
           coverage run --branch -m pytest
           coverage xml
+
       - name: Upload results to Codecov
         if: ${{ (github.event_name == 'push' && github.ref == 'refs/heads/main') || (github.event_name == 'pull_request' && github.base_ref == 'main') }}
         uses: codecov/codecov-action@v5
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
           slug: AI-Smarties/back
 
-
   docker:
     if: ${{ (github.event_name == 'push' && github.ref == 'refs/heads/main') || (github.event_name == 'pull_request' && github.base_ref == 'main') }}
     runs-on: ubuntu-latest
     needs: tests
     steps:
       - name: Checkout
         uses: actions/checkout@v6
+
       - name: Set up Docker Buildx
         uses: docker/setup-buildx-action@v3
+
       - name: Login to Docker Hub
         if: ${{ github.event_name == 'push' && github.ref == 'refs/heads/main' }}
         uses: docker/login-action@v3
         with:
           username: ${{ vars.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_TOKEN }}
+
       - name: Build image (and push)
         uses: docker/build-push-action@v6
         with:

requirements.txt

@@ -12,3 +12,4 @@ pgvector
 
 vertexai
 google-genai
+tzdata

scripts/corridor_talk_transcript.txt

@@ -0,0 +1,21 @@
+Hey Tom, got a sec? I wanted to catch up before the standup.
+
+Sure, just grabbing coffee. Want some?
+
+No thanks, I'm already on my third cup today. So, um, I was thinking about the Jenkins pipeline issue — it's been flaky again this week.
+
+Yeah I saw that. Probably just the test timeouts, nothing serious. I'll look at it Thursday.
+
+Cool. Oh — also, I talked to the client this morning. Big news actually: they're moving the launch from Q3 to May 15th.
+
+Wait, seriously? That's six weeks earlier.
+
+Yeah. And they want the analytics dashboard included in the initial release now, not post-launch.
+
+Okay... that changes everything. We're going to need another backend dev at minimum. I can't promise the dashboard by May 15th with the current team.
+
+Agreed. I'll escalate to Lisa today and request the hire. We're officially committing to May 15th with the expanded scope though — that's confirmed with the client.
+
+Got it. I'll reshuffle the sprint after standup. Oh — did you see the parking memo? They're changing the visitor spots again.
+
+Ha, no I didn't. Okay, see you in standup.

src/db_utils.py

@@ -4,6 +4,7 @@
 from vertexai.language_models import TextEmbeddingModel
 import google
 from sqlalchemy import select
+from sqlalchemy.orm import joinedload
 from db import sessionlocal
 from models import Conversation, Vector, Category, EMBEDDING_DIMENSIONS
 
@@ -47,16 +48,23 @@ def get_vectors():
     with sessionlocal() as session:
         return session.scalars(select(Vector)).all()
 
-def search_vectors(text, limit=1):
+def search_vectors(text, limit=1, max_distance=0.5):
     if not EMBEDDING_MODEL:
         load_embedding_model()
     embedding = EMBEDDING_MODEL.get_embeddings(
         [text],
         output_dimensionality=EMBEDDING_DIMENSIONS,
     )[0].values
     with sessionlocal() as session:
+        # how "relevant" the query response should be on scale of 0-2 (float)
+        # 0 = identical, 1 = unrelated, 2 = opposite
+        distance = Vector.embedding.cosine_distance(embedding)
         return session.scalars(
-            select(Vector).order_by(Vector.embedding.cosine_distance(embedding)).limit(limit)
+            select(Vector)
+            .options(joinedload(Vector.conversation))
+            .where(distance < max_distance)
+            .order_by(distance)
+            .limit(limit)
         ).all()
 
 def create_conversation(name, summary=None, cat_id=None, timestamp=None):