add in rollups calculator

db/db_utils.py

@@ -53,6 +53,29 @@ def initialize_db(conn):
             updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
         )
     ''')
+    cursor.execute('''
+        CREATE TABLE IF NOT EXISTS donations (
+            run_key TEXT,
+            run_row INTEGER,
+            year TEXT,
+            from TEXT,
+            to TEXT,
+            total_emissions_reduction REAL,
+            created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
+            updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
+        )
+    ''')
+    cursor.execute('''
+        CREATE TABLE IF NOT EXISTS rollups (
+            run_key TEXT,
+            year TEXT,
+            donations_discount REAL,
+            total_emissions_reduction_pre REAL,
+            total_emissions_reduction REAL,
+            created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
+            updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
+        )
+    ''')
     cursor.execute('''
         CREATE TABLE IF NOT EXISTS results (
             id BIGSERIAL PRIMARY KEY,
@@ -86,6 +109,8 @@ def initialize_db(conn):
         CREATE INDEX IF NOT EXISTS idx_cleaned_word ON mappings(cleaned_word);
         CREATE INDEX IF NOT EXISTS idx_dictionary_word ON mappings(dictionary_word);
         CREATE INDEX IF NOT EXISTS idx_description ON dictionary(description);
+        CREATE UNIQUE INDEX IF NOT EXISTS run_row_run_key_uniq ON results(run_key text_ops,run_row int4_ops);
+        CREATE UNIQUE INDEX IF NOT EXISTS run_row_run_key_uniq ON donations(run_key text_ops,run_row int4_ops);
     ''')
 
     conn.commit()

rollups.py

@@ -0,0 +1,73 @@
+from db_utils import get_connection
+import re
+
+# Function to extract the year
+def extract_year(string):
+    match = re.search(r'\b(19|20)\d{2}\b', string)
+    if match:
+        return match.group()
+    else:
+        return None
+
+db_conn = get_connection()
+db_cursor = db_conn.cursor()
+
+db_cursor.execute("SELECT name from donors")
+donor_results = db_cursor.fetchall()
+list_of_valid_donors = [row[0] for row in donor_results]
+
+
+db_cursor.execute("SELECT distinct(run_key) FROM results")
+result_rows = db_cursor.fetchall()
+
+run_keys_and_years = []
+for row in result_rows:
+    run_key = row[0]
+    # calculate the total_emissions_reductions of items that were donated
+    # or received as donations
+
+    # to do this, we parse the 4 digit year from the run_key
+    year = extract_year(run_key)
+    run_keys_and_years.append((run_key, year))
+
+
+for run_key, year in run_keys_and_years:
+    donations_to = 0
+    donations_from = 0
+
+    # find all run_keys for the same year
+    run_keys_in_year = [run_key for run_key, y in run_keys_and_years if y == year]
+
+    # find all donations from the current donor
+    db_cursor.execute("SELECT sum(total_emissions_reduction) FROM results WHERE run_key IN %s and run_key != %s and donor in %s", (tuple(run_keys_in_year), run_key, tuple(list_of_valid_donors)))
+    result = db_cursor.fetchone()
+    if result:
+        donations_from = result[0]
+    
+    # divide the donations_from by 2 because we attribute the emissions reductions to both the donor and the recipient
+    donations_from = donations_from / 2
+    
+    # find all donations to the current donor
+    db_cursor.execute("SELECT sum(total_emissions_reduction) FROM results WHERE run_key = %s donor in %s", (run_key, tuple(list_of_valid_donors)))
+    result = db_cursor.fetchone()
+    if result:
+        donations_to = result[0]
+    
+    # divide the donations_to by 2 because we attribute the emissions reductions to both the donor and the recipient
+    donations_to = donations_to / 2
+
+    # calculate the total_donations
+    donations = donations_to + donations_from
+
+    # calculate total_emissions_reduction_pre
+    db_cursor.execute("SELECT sum(total_emissions_reduction) FROM results WHERE run_key = %s", (run_key,))
+    result = db_cursor.fetchone()
+    if result:
+        total_emissions_reduction_pre = result[0]
+    else:
+        total_emissions_reduction_pre = 0
+
+    total_emissions_reduction = total_emissions_reduction_pre - donations
+
+    # store this data in the rollups table
+    db_cursor.execute("INSERT INTO rollups (run_key, year, donations_discount, total_emissions_reduction_pre, total_emissions_reduction) VALUES (%s, %s, %s, %s, %s)", (run_key, year, donations, total_emissions_reduction_pre, total_emissions_reduction, ))