From 6fe474a2f84d7c3a03bfa56e9d91945ed05692bc Mon Sep 17 00:00:00 2001
From: Joshua Kwan <joshua.m.kwan@gmail.com>
Date: Mon, 13 Nov 2023 07:13:31 -0500
Subject: [PATCH] Aggregate failures in console output (#98)

Failures are grouped by the check name, description, and aggregate the
distinct values and rows where those failing values occur

Signed-off-by: Joshua Kwan <joshk@ternary.app>
---
 .../outputter/outputter_console.py            | 28 ++++++++++++++++++-
 1 file changed, 27 insertions(+), 1 deletion(-)

diff --git a/focus_validator/outputter/outputter_console.py b/focus_validator/outputter/outputter_console.py
index 759f46a..9c6260d 100644
--- a/focus_validator/outputter/outputter_console.py
+++ b/focus_validator/outputter/outputter_console.py
@@ -1,3 +1,4 @@
+import math
 import pandas as pd
 from tabulate import tabulate
 
@@ -27,6 +28,7 @@ def __restructure_check_list__(result_set: ValidationResult):
                 }
             )
             rows.append(row_obj)
+
         df = pd.DataFrame(rows)
         df.rename(
             columns={
@@ -59,11 +61,35 @@ def write(self, result_set: ValidationResult):
         print(tabulate(checklist, headers="keys", tablefmt="psql"))
 
         if result_set.failure_cases is not None:
+            aggregated_failures = result_set.failure_cases.groupby(by=['Check Name', 'Column', 'Description'], as_index=False).aggregate(lambda x: maybe_collapse_range(x.unique().tolist()))
+
             print("Checks summary:")
             print(
                 tabulate(
-                    tabular_data=result_set.failure_cases,  # type: ignore
+                    tabular_data=aggregated_failures, # type: ignore
                     headers="keys",
                     tablefmt="psql",
                 )
             )
+
+def maybe_collapse_range(l):
+    start = None
+    i = None
+    collapsed = []
+    for n in sorted(l):
+        if not isinstance(n, int) and not (isinstance(n, float) and not math.isnan(n)):
+            return l
+        elif i is None:
+            start = i = n
+        elif n == i + 1:
+            i = n
+        elif i:
+            if i == start: collapsed.append(f'{int(start)}')
+            else: collapsed.append(f'{int(start)}-{int(i)}')
+            start = i = n
+
+    if start is not None:
+        if i == start: collapsed.append(int(start))
+        else: collapsed.append(f'{int(start)}-{int(i)}')
+
+    return collapsed