From 60d5bc686b4739de3835c567d4e8ada02174612d Mon Sep 17 00:00:00 2001
From: Simon Roberts <lyricnz@gmail.com>
Date: Sun, 24 Mar 2024 13:05:04 +1100
Subject: [PATCH 1/4] Add shebang and +x to make adhoc_tools.py directly
 executable

---
 code/adhoc_tools.py | 1 +
 1 file changed, 1 insertion(+)
 mode change 100644 => 100755 code/adhoc_tools.py

diff --git a/code/adhoc_tools.py b/code/adhoc_tools.py
old mode 100644
new mode 100755
index 84705a0336..9c07401562
--- a/code/adhoc_tools.py
+++ b/code/adhoc_tools.py
@@ -1,3 +1,4 @@
+#!/usr/bin/env python3
 import argparse
 import csv
 import glob

From d2bebe221475860693c52b735259de8f2e964e7b Mon Sep 17 00:00:00 2001
From: Simon Roberts <lyricnz@gmail.com>
Date: Sun, 24 Mar 2024 13:12:12 +1100
Subject: [PATCH 2/4] Add progress to generate_all_suburbs_nbn_tallies

---
 code/adhoc_tools.py | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/code/adhoc_tools.py b/code/adhoc_tools.py
index 9c07401562..67166d731c 100755
--- a/code/adhoc_tools.py
+++ b/code/adhoc_tools.py
@@ -239,7 +239,11 @@ def generate_all_suburbs_nbn_tallies():
     """Create a file containing a tally of all suburbs by property (tech, upgrade, etc)"""
     exclude_properties = {"name", "locID", "gnaf_pid"}
     tallies = {}  # property-name -> Counter()
-    for file in glob.glob("results/**/*.geojson"):
+    filenames = glob.glob("results/**/*.geojson")
+    for n, file in enumerate(filenames):
+        if n % 100 == 0:
+            utils.print_progress_bar(n, len(filenames), prefix="Progress:", suffix="Complete", length=50)
+
         for feature in utils.read_json_file(file)["features"]:
             for prop, value in feature["properties"].items():
                 if prop not in exclude_properties:

From 6fc02e3570b18a94f4c1ea96c21a377b76330383 Mon Sep 17 00:00:00 2001
From: Simon Roberts <lyricnz@gmail.com>
Date: Sun, 24 Mar 2024 13:36:31 +1100
Subject: [PATCH 3/4] sort tallies by frequency, except
 'target_eligibility_quarter' which is sorted by date

---
 code/adhoc_tools.py | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/code/adhoc_tools.py b/code/adhoc_tools.py
index 67166d731c..d87bfdd40c 100755
--- a/code/adhoc_tools.py
+++ b/code/adhoc_tools.py
@@ -251,6 +251,16 @@ def generate_all_suburbs_nbn_tallies():
                         tallies[prop] = Counter()
                     tallies[prop][value] += 1
 
+    def _parse_quarter(item: tuple[str, int]):
+        """Parse a quarter string into a datetime object.  If NA, return epoch."""
+        return datetime.fromtimestamp(0) if item[0] == "NA" else datetime.strptime(item[0], "%b %Y")
+
+    # sort tallies by frequency, except 'target_eligibility_quarter' which is sorted by date
+    tallies = {
+        k: OrderedDict(sorted(v.items(), key=_parse_quarter) if k == "target_eligibility_quarter" else v.most_common())
+        for k, v in tallies.items()
+    }
+
     # Add percentages and missing items
     total_count = sum(tallies["tech"].values())  # everything has a tech+NULL
     tallies["percent"] = {}

From dac1b99e3d8eb295befeb84c73d02234e44aeb00 Mon Sep 17 00:00:00 2001
From: Simon Roberts <lyricnz@gmail.com>
Date: Sun, 24 Mar 2024 13:46:34 +1100
Subject: [PATCH 4/4] Handle bad date formats better

---
 code/adhoc_tools.py | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/code/adhoc_tools.py b/code/adhoc_tools.py
index d87bfdd40c..ff5970d5a5 100755
--- a/code/adhoc_tools.py
+++ b/code/adhoc_tools.py
@@ -253,7 +253,10 @@ def generate_all_suburbs_nbn_tallies():
 
     def _parse_quarter(item: tuple[str, int]):
         """Parse a quarter string into a datetime object.  If NA, return epoch."""
-        return datetime.fromtimestamp(0) if item[0] == "NA" else datetime.strptime(item[0], "%b %Y")
+        try:
+            return datetime.strptime(item[0], "%b %Y")
+        except ValueError:
+            return datetime.fromtimestamp(0)
 
     # sort tallies by frequency, except 'target_eligibility_quarter' which is sorted by date
     tallies = {