Merge pull request #29 from RobotSail/feedback-updates

add initial feedback code

app/app.py

@@ -6,6 +6,7 @@
     eval_using_model,
     indicate_key_presence,
 )
+from feedback import store_feedback
 import os
 import streamlit as st
 import logging
@@ -15,6 +16,8 @@
 from sklearn.metrics.pairwise import cosine_similarity
 from textstat import textstat
 import os
+from streamlit_feedback import streamlit_feedback
+
 
 # Set theme, title, and icon
 st.set_page_config(page_title="API Docs Generator", page_icon="📄", layout="wide")
@@ -219,6 +222,7 @@ def main(prompt_success: bool, prompt_diff: int, actual_doc: str):
             top_p,
             GENAI_KEY(),
         )
+
     col1, col2, col3 = st.columns([1.5, 1.5, 0.5])
 
     with col1:
@@ -308,3 +312,11 @@ def main(prompt_success: bool, prompt_diff: int, actual_doc: str):
         main(prompt_success, prompt_diff, actual_doc)
     else:
         main(True, True, actual_doc)
+
+# generate the feedback section now
+streamlit_feedback(
+    feedback_type="thumbs",
+    on_submit=store_feedback,
+    optional_text_label="Please tell us how we could make this more useful",
+    align="flex-start",
+)

app/feedback.py

@@ -0,0 +1,84 @@
+import logging
+import sqlite3
+import os
+import time
+
+logger = logging.getLogger(__name__)
+
+
+def get_db_file() -> str:
+    """
+    Returns either the database location that was provided
+    via the FEEDBACK_DB environment variable, or returns
+    a default location.
+    """
+    db_file = os.getenv("FEEDBACK_DB")
+    if not db_file:
+        db_file = "/tmp/feedback.db"
+        logger.warning(
+            "No feedback database file specified, storing feedback in default location: '%s'",
+            db_file,
+        )
+    return db_file
+
+
+def create_connection() -> sqlite3.Connection | None:
+    """
+    Creates connection with the Sqlite database.
+    """
+
+    db_file = get_db_file()
+    conn = None
+    try:
+        conn = sqlite3.connect(db_file)
+    except sqlite3.Error as e:
+        logger.error(e)
+    return conn
+
+
+def create_table(conn: sqlite3.Connection):
+    """
+    Creates the feedback table if it does not exist.
+    """
+    try:
+        sql_create_feedback_table = """ CREATE TABLE IF NOT EXISTS feedback (
+                                            id integer PRIMARY KEY AUTOINCREMENT,
+                                            type text NOT NULL,
+                                            score text NOT NULL,
+                                            text text,
+                                            timestamp text NOT NULL
+                                        ); """
+        c = conn.cursor()
+        c.execute(sql_create_feedback_table)
+    except sqlite3.Error as e:
+        logger.error(e)
+
+
+def store_feedback(feedback: dict) -> None:
+    """
+    Stores feedback from the provided feedback dict, which
+    comes from the streamlit-feedback component.
+
+    If the database does not exist, it will be created.
+    """
+    logger.info("Got feedback: %s", str(feedback))
+    conn = create_connection()
+    if conn is not None:
+        create_table(conn)
+        sql = """ INSERT INTO feedback(type,score,text,timestamp)
+                    VALUES(?,?,?,?) """
+        cur = conn.cursor()
+        logger.debug("Storing feedback in database.")
+        cur.execute(
+            sql,
+            (
+                feedback["type"],
+                feedback["score"],
+                feedback.get("text", ""),
+                time.ctime(),
+            ),
+        )
+        conn.commit()
+        logger.debug("Feedback stored in database.")
+    else:
+        logger.error("Error! cannot create the database connection.")

app/requirements.txt

@@ -10,3 +10,4 @@ py-readability-metrics
 openai
 textstat
 scikit-learn
+streamlit-feedback

app/resources/deployment.yaml

@@ -14,6 +14,10 @@ spec:
     spec:
       imagePullSecrets:
       - name: dockerhub-secret
+      volumes:
+        - name: user-feedback-db
+          persistentVolumeClaim:
+            claimName: user-feedback-db
       containers:
       - name: api-doc-container
         image: docker.io/ochatterjee/api-docs-gen:v0.1.0
@@ -29,6 +33,8 @@ spec:
               secretKeyRef:
                 name: api-doc-secret
                 key: OPENAI_API_KEY
+          - name: FEEDBACK_DB
+            value: /mnt/user-feedback-db/feedback.db
         ports:
         - containerPort: 8501
         resources:
@@ -38,3 +44,6 @@ spec:
           requests:
             memory: "200Mi"
             cpu: "250m"
+        volumeMounts:
+          - mountPath: "/mnt/user-feedback-db"
+            name: user-feedback-db

app/resources/pvc.yaml

@@ -0,0 +1,11 @@
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: user-feedback-db
+spec:
+  accessModes:
+    - ReadWriteOnce
+  resources:
+    requests:
+      storage: 1Gi

app/utils.py

@@ -1,13 +1,13 @@
+import json
+import os
+from openai import OpenAI
 from genai import Credentials, Client
 from genai.text.generation import TextGenerationParameters
 from genai.text.tokenization import (
     TextTokenizationParameters,
     TextTokenizationReturnOptions,
     TextTokenizationCreateResults,
 )
-import json
-from openai import OpenAI
-import os
 
 
 def generate_prompt(

notebooks/data-collection/user-feedback.ipynb

@@ -0,0 +1,134 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Analyzing Feedback\n",
+    "\n",
+    "This Jupyter notebook demonstrates how to extract\n",
+    "user feedback from the database we created earlier.\n",
+    "\n",
+    "The data is stored in a table which was created with the\n",
+    "following schema:\n",
+    "\n",
+    "```sql\n",
+    "CREATE TABLE IF NOT EXISTS feedback (\n",
+    "\tid integer PRIMARY KEY AUTOINCREMENT,\n",
+    "\ttype text NOT NULL,\n",
+    "\tscore text NOT NULL,\n",
+    "\ttext text,\n",
+    "\ttimestamp text NOT NULL\n",
+    ");\n",
+    "```"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import sqlite3\n",
+    "import os"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def get_db_file() -> str:\n",
+    "    db_file = os.getenv(\"FEEDBACK_DB\")\n",
+    "    if not db_file:\n",
+    "        db_file = \"/tmp/feedback.db\"\n",
+    "    return db_file\n",
+    "\n",
+    "\n",
+    "def create_connection() -> sqlite3.Connection | None:\n",
+    "    \"\"\"\n",
+    "    Creates connection with the Sqlite database.\n",
+    "    \"\"\"\n",
+    "\n",
+    "    db_file = get_db_file()\n",
+    "    conn = None\n",
+    "    try:\n",
+    "        conn = sqlite3.connect(db_file)\n",
+    "    except sqlite3.Error as e:\n",
+    "        print(e)\n",
+    "    return conn"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "conn = create_connection()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "(4, 'thumbs', '👍', 'The API response is already perfect 👍', 'Tue Jan 16 17:22:09 2024')\n",
+      "(5, 'thumbs', '👎', 'The API was way too long', 'Tue Jan 16 17:22:50 2024')\n"
+     ]
+    }
+   ],
+   "source": [
+    "conn = create_connection()\n",
+    "\n",
+    "# Execute the SELECT * query\n",
+    "cursor = conn.cursor()\n",
+    "cursor.execute(\"SELECT * FROM feedback\")\n",
+    "\n",
+    "# Fetch all rows from the result set\n",
+    "rows = cursor.fetchall()\n",
+    "\n",
+    "# Print the rows\n",
+    "for row in rows:\n",
+    "    print(row)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Close the cursor and connection\n",
+    "cursor.close()\n",
+    "conn.close()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "venv",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.1"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}