commit dd7851c1143dc5c03848030a3f0c49deec6c8461
Author: Paul Trowbridge <paul@hptrow.me>
Date:   Mon Mar 27 11:27:12 2023 -0400

    python script to read wav files, call openai api, and write contents to database

diff --git a/.env_sample b/.env_sample
new file mode 100644
index 0000000..01c7ffe
--- /dev/null
+++ b/.env_sample
@@ -0,0 +1,5 @@
+DB_HOST=hostname
+DB_NAME=database_name
+DB_USER=username
+DB_PASSWORD=password
+OPENAI_API_KEY=sk-token
diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..752659b
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,3 @@
+.env
+*.swp
+curl.sh
diff --git a/ddl.pg.sql b/ddl.pg.sql
new file mode 100644
index 0000000..173e918
--- /dev/null
+++ b/ddl.pg.sql
@@ -0,0 +1,10 @@
+CREATE TABLE IF NOT EXISTS 
+rlarp.thirtysec (
+    filename    text
+    ,mdate      date
+    ,message    jsonb
+);
+
+ALTER TABLE rlarp.thirtysec ADD PRIMARY KEY (filename);
+
+GRANT ALL ON TABLE rlarp.thirtysec TO report;
diff --git a/transcribe.py b/transcribe.py
new file mode 100644
index 0000000..e6cda5d
--- /dev/null
+++ b/transcribe.py
@@ -0,0 +1,68 @@
+import requests
+import psycopg2
+import json
+import os
+from dotenv import load_dotenv
+
+load_dotenv()
+db_host = os.getenv('DB_HOST')
+db_name = os.getenv('DB_NAME')
+db_user = os.getenv('DB_USER')
+db_password = os.getenv('DB_PASSWORD')
+openai_api_key = os.getenv('OPENAI_API_KEY')
+
+# Set up the database connection
+conn = psycopg2.connect(
+     host=       db_host
+    ,database=   db_name
+    ,user=       db_user
+    ,password=   db_password
+    ,connect_timeout = 120
+)
+
+# Define the API endpoint and headers
+url = 'https://api.openai.com/v1/audio/translations'
+headers = {
+    'Authorization': f'Bearer {openai_api_key}'
+}
+params = {
+    'model': 'whisper-1',
+    'response_format': 'vtt'
+}
+data = {
+    'model': 'whisper-1'
+}
+
+dir_path = '//mnt/c/Users/PTrowbridge/Downloads/trans'
+# Define the audio file to be transcribed
+for file_name in os.listdir(dir_path):
+    if file_name.endswith('.wav'):
+        file_path = os.path.join(dir_path, file_name)
+        file_name = os.path.basename(file_path)
+        file_date = file_name[:10]
+
+        # Check if there is a row in the database with a matching filename
+        cur = conn.cursor()
+        cur.execute("SELECT COUNT(*) FROM rlarp.thirtysec WHERE filename = %s", (file_name,))
+        count = cur.fetchone()[0]
+        cur.close()
+        if count > 0:
+            print(f"Skipping {file_name} (already processed)")
+            continue
+
+        # Send the transcription request and retrieve the results
+        audio_file = open(file_path, 'rb')
+        response = requests.post(url, headers=headers, params=params, data=data, files={'file': audio_file})
+        transcript = response.text
+
+        print(response.text)
+
+        # Insert the JSON summary into the database
+        cur = conn.cursor()
+        cur.execute("INSERT INTO rlarp.thirtysec (filename, mdate, message) VALUES (%s, %s, %s);", (file_name, file_date, response.text))
+        conn.commit()
+        cur.close()
+
+#close db connection
+conn.close()
+