From 7191358b8d058ec549aed4e1aec8b29696dccad9 Mon Sep 17 00:00:00 2001
From: Seth Call <seth@jamkazam.com>
Date: Sat, 8 Feb 2025 23:02:32 -0600
Subject: [PATCH] sluggarize script

---
 web/script/py/requirements.txt        |  1 +
 web/script/py/sluggarize_jamtracks.py | 94 +++++++++++++++++++++++++++
 2 files changed, 95 insertions(+)
 create mode 100644 web/script/py/requirements.txt
 create mode 100644 web/script/py/sluggarize_jamtracks.py

diff --git a/web/script/py/requirements.txt b/web/script/py/requirements.txt
new file mode 100644
index 000000000..658130bb2
--- /dev/null
+++ b/web/script/py/requirements.txt
@@ -0,0 +1 @@
+psycopg2
diff --git a/web/script/py/sluggarize_jamtracks.py b/web/script/py/sluggarize_jamtracks.py
new file mode 100644
index 000000000..3210df803
--- /dev/null
+++ b/web/script/py/sluggarize_jamtracks.py
@@ -0,0 +1,94 @@
+import psycopg2
+import re
+import argparse
+import os
+from urllib.parse import quote
+
+# Database connection settings from environment variables
+# Function to construct DB_CONFIG dynamically based on the environment
+def get_db_config(env):
+    env = env.upper()
+    if env not in ["DEV", "STAGING", "PROD"]:
+        raise ValueError("Invalid environment. Choose from dev, staging, or prod.")
+
+    return {
+        "dbname": os.getenv(f"JAM_DB_{env}_DBNAME", "jam"),
+        "user": os.getenv(f"JAM_DB_{env}_USER", "jam"),
+        "password": os.getenv(f"JAM_DB_{env}_PASSWORD", ""),
+        "host": os.getenv(f"JAM_DB_{env}_HOST", "127.0.0.1"),
+        "port": os.getenv(f"JAM_DB_{env}_PORT", "5432")
+    }
+
+
+import unicodedata
+def slugify(text):
+    """Convert a string into a URL-friendly slug."""
+    text = text.lower()
+    text = unicodedata.normalize('NFKD', text).encode('ascii', 'ignore').decode('utf-8')
+    text = re.sub(r"[\s]+", "-", text)  # Replace whitespace with '-'
+    text = re.sub(r"[\:]+", "-", text)  # Replace : with '-'
+    text = re.sub(r"[&\+]+", "and", text)  # Replace & with and
+    text = re.sub(r"['\",!?\(\)\=\#\+\*\.]", "", text)  # Remove quotes, commas, ?, and !
+    text = re.sub(r"[\-]+", "-", text)  # Replace consecutive whitespace with '-'
+
+    #text = re.sub(r"[^a-z0-9-]", "", text)  # Remove all non-alphanumeric except '-'
+    if text.startswith("-") or text.endswith("-"):
+        text = text.strip("-")
+    return text
+
+def update_slugs(env, dry_run=False):
+    """Fetch records, generate slugs, and update the table."""
+    try:
+        db_config_for_env = get_db_config(env)
+        conn = psycopg2.connect(**db_config_for_env)
+        cursor = conn.cursor()
+        
+        # Fetch all records
+        cursor.execute("SELECT id, original_artist, name FROM jam_tracks;")
+        rows = cursor.fetchall()
+        
+        for row in rows:
+            track_id, original_artist, name = row
+            #print(f"{track_id} : {original_artist} : {name}")
+            artist_slug = slugify(original_artist) if original_artist else ""
+            name_slug = slugify(name) if name else ""
+            
+            # Validate percent encoding
+            encoded_artist_slug = quote(artist_slug)
+            encoded_name_slug = quote(name_slug)
+            
+            if encoded_artist_slug != artist_slug:
+                print(f"Warning: Slug mismatch for original_artist '{original_artist}' -> '{artist_slug}'")
+            if encoded_name_slug != name_slug:
+                print(f"Warning: Slug mismatch for name '{name}' -> '{name_slug}'")
+
+            #print(f"{artist_slug}\n{name_slug}")
+            if not dry_run:
+                # Update database with new slugs
+                cursor.execute(
+                    "UPDATE jam_tracks SET original_artist_slug = %s, name_slug = %s WHERE id = %s;",
+                    (artist_slug, name_slug, track_id)
+                )
+        
+        if not dry_run:
+            conn.commit()
+            print(f"Updated {cursor.rowcount} rows.")
+    
+    except Exception as e:
+        print("Error:", e)
+    
+    finally:
+        if cursor:
+            cursor.close()
+        if conn:
+            conn.close()
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--dry-run", action="store_true", help="Run in dry-run mode without updating the database")
+    parser.add_argument("--env", choices=["dev", "staging", "prod"], default="dev", help="Specify the environment (dev, staging, prod)")
+
+    args = parser.parse_args()
+    
+    update_slugs(args.env, dry_run=args.dry_run)
+