From d193835301f8ef41ea455ff1a135e0c409623324 Mon Sep 17 00:00:00 2001
From: ds <63077097+dsmedia@users.noreply.github.com>
Date: Tue, 23 Jul 2024 07:15:42 -0400
Subject: [PATCH] Add source for budget.json and validation script Partially
 addresses #15

---
 SOURCES.md                         |  3 +++
 scripts/validate_budget_dataset.py | 25 +++++++++++++++++++++++++
 2 files changed, 28 insertions(+)
 create mode 100644 scripts/validate_budget_dataset.py

diff --git a/SOURCES.md b/SOURCES.md
index bad6739..d83d3f9 100644
--- a/SOURCES.md
+++ b/SOURCES.md
@@ -26,6 +26,9 @@ http://wildlife.faa.gov
 
 ## `budget.json`
 
+Source: Office of Management and Budget (U.S.)
+[Budget FY 2016 - Receipts](https://www.govinfo.gov/app/details/BUDGET-2016-DB/BUDGET-2016-DB-3)
+
 ## `budgets.json`
 
 ## `burtin.json`
diff --git a/scripts/validate_budget_dataset.py b/scripts/validate_budget_dataset.py
new file mode 100644
index 0000000..8fbee5f
--- /dev/null
+++ b/scripts/validate_budget_dataset.py
@@ -0,0 +1,25 @@
+import pandas as pd
+
+def validate_budget_data():
+    """
+    Validates that the budget data in the vega-datasets repository
+    matches the original source from the U.S. Government Publishing Office.
+    """
+    source_url = "https://www.govinfo.gov/content/pkg/BUDGET-2016-DB/xls/BUDGET-2016-DB-3.xls" # https://www.govinfo.gov/app/details/BUDGET-2016-DB/context
+    vega_url = "https://raw.githubusercontent.com/vega/vega-datasets/05fcb7c07b1d76206856e75129fc1e79dc61735c/data/budget.json" # 2015-10-15 Commit
+
+    source_df = pd.read_excel(source_url)
+    vega_df = pd.read_json(vega_url)
+
+    # Format numeric columns and 'TQ' with commas
+    numeric_cols = [col for col in source_df.columns if col.isdigit()] + ['TQ']
+    source_df[numeric_cols] = source_df[numeric_cols].apply(lambda col: col.map(lambda x: f"{x:,}" if pd.notnull(x) else x))
+
+    try:
+        pd.testing.assert_frame_equal(source_df, vega_df)
+        print("The DataFrames are identical.")
+    except AssertionError as e:
+        print("The DataFrames are not identical. Differences:", e)
+
+if __name__ == "__main__":
+    validate_budget_data()
\ No newline at end of file