create and insert google pinyin with abbr of pinyin to sqlite3 db use…

… Python 3.12.2

dictionary/sqlite/create_pinyin_db_schema.py

@@ -0,0 +1,44 @@
+import sqlite3
+
+# Path to the SQLite database file
+db_file = 'pinyin_data.sqlite3'
+
+# Connect to the database. This will create the file if it doesn't exist.
+conn = sqlite3.connect(db_file)
+
+# Create a cursor object using the connection
+cur = conn.cursor()
+
+# SQL statement to create the pinyin_data table
+# hz hanzhi
+# py pinyin
+# abbr abbreviation
+# freq frequency
+create_table_sql = """
+CREATE TABLE IF NOT EXISTS pinyin_data (
+    id INTEGER PRIMARY KEY AUTOINCREMENT,
+    hz TEXT NOT NULL,
+    py TEXT NOT NULL,
+    abbr TEXT NOT NULL,
+    freq REAL NOT NULL
+);
+"""
+
+# Execute the SQL statement to create the table
+cur.execute(create_table_sql)
+
+# SQL statements to create indexes on the pinyin and abbreviation columns
+create_index_pinyin_sql = "CREATE INDEX IF NOT EXISTS idx_pinyin ON pinyin_data(py);"
+create_index_abbreviation_sql = "CREATE INDEX IF NOT EXISTS idx_abbr ON pinyin_data(abbr);"
+
+# Execute the SQL statements to create the indexes
+cur.execute(create_index_pinyin_sql)
+cur.execute(create_index_abbreviation_sql)
+
+# Commit the changes
+conn.commit()
+
+# Close the connection
+conn.close()
+
+print("Database, table, and indexes created successfully.")

dictionary/sqlite/insert_pinyin_to_sqlite_db.py

@@ -0,0 +1,28 @@
+import sqlite3
+
+# Connect to your SQLite database
+conn = sqlite3.connect('pinyin_data.sqlite3')
+c = conn.cursor()
+
+# Open and read your data file
+with open('../google_pinyin_rawdict_utf16_65105_freq.txt', 'r', encoding='utf-16') as file:
+    for line in file:
+        # Split the line into components
+        parts = line.strip().split(' ')
+
+        # Omit the '0' and reconstruct the line if necessary
+        # Assuming the format is consistent and '0' always appears at the third position
+        if parts[2] == '0':
+            phrase = parts[0]
+            score = parts[1]
+            pinyin = ''.join(parts[3:])  # Join the remaining parts as the pinyin
+            abbreviation = ''.join([p[0] for p in parts[3:]])  # Create the abbreviation from the pinyin parts
+            print(phrase, pinyin, abbreviation, float(score))
+            # Execute the insert command
+            c.execute('INSERT INTO pinyin_data (hz, py, abbr, freq) VALUES (?, ?, ?, ?)', 
+                (phrase, pinyin, abbreviation, float(score)))
+
+
+# Commit the changes and close the connection
+conn.commit()
+conn.close()