From 960c4b76193d96a447b7638725490b98993fd0b3 Mon Sep 17 00:00:00 2001
From: alexobaseki <alex@pluralpolicy.com>
Date: Wed, 25 Dec 2024 11:59:22 -0500
Subject: [PATCH 1/3] Clean up phone

---
 openstates/cli/convert_us.py | 17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)

diff --git a/openstates/cli/convert_us.py b/openstates/cli/convert_us.py
index 5ab1208c1..449fefe69 100644
--- a/openstates/cli/convert_us.py
+++ b/openstates/cli/convert_us.py
@@ -1,3 +1,4 @@
+import re
 import typing
 import uuid
 from collections import defaultdict
@@ -28,7 +29,22 @@ def make_org_id(id_: str) -> str:
     return "ocd-organization/" + str(uuid.uuid5(US_UUID_NAMESPACE, id_))
 
 
+def sanitize_phone(phone: str) -> str:
+    """Remove trail text, toll-free phone number or N/A"""
+    if phone.lower() in ["n/a", "same as above"]:
+        return ""
+
+    pattern = r"\((\d{3})\)\s*(\d{3})-(\d{4})"
+    match = re.search(pattern, phone)
+    if match:
+        # Format the first matched number as XXX-XXX-XXXX
+        formatted_number = f"{match.group(1)}-{match.group(2)}-{match.group(3)}"
+        return formatted_number
+    return phone
+
+
 def _fix_bad_dashes(phone: str) -> str:
+    phone = sanitize_phone(phone)
     return phone.replace("–", "-")
 
 
@@ -43,7 +59,6 @@ def get_district_offices() -> defaultdict[str, list[Office]]:
                 if office.get("suite"):
                     address += " " + office["suite"]
                 address += f"; {office['city']}, {office['state']} {office['zip']}"
-
             district_offices[entry["id"]["bioguide"]].append(
                 Office(
                     classification="district",

From f80411f7729e34576e17b7f8e9bf3014c3c4ff5b Mon Sep 17 00:00:00 2001
From: alexobaseki <alex@pluralpolicy.com>
Date: Wed, 25 Dec 2024 12:10:23 -0500
Subject: [PATCH 2/3] add some comments

---
 openstates/cli/convert_us.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/openstates/cli/convert_us.py b/openstates/cli/convert_us.py
index 449fefe69..c0fe433da 100644
--- a/openstates/cli/convert_us.py
+++ b/openstates/cli/convert_us.py
@@ -33,7 +33,7 @@ def sanitize_phone(phone: str) -> str:
     """Remove trail text, toll-free phone number or N/A"""
     if phone.lower() in ["n/a", "same as above"]:
         return ""
-
+    # Some phone might appear like (123) 456 7890
     pattern = r"\((\d{3})\)\s*(\d{3})-(\d{4})"
     match = re.search(pattern, phone)
     if match:
@@ -59,6 +59,7 @@ def get_district_offices() -> defaultdict[str, list[Office]]:
                 if office.get("suite"):
                     address += " " + office["suite"]
                 address += f"; {office['city']}, {office['state']} {office['zip']}"
+
             district_offices[entry["id"]["bioguide"]].append(
                 Office(
                     classification="district",

From 9a67e1e6656c3fa9187a891064d1706a4ebf3119 Mon Sep 17 00:00:00 2001
From: alexobaseki <alex@pluralpolicy.com>
Date: Fri, 27 Dec 2024 15:46:13 -0500
Subject: [PATCH 3/3] Bump version to 6.20.13

---
 CHANGELOG.md   | 4 ++++
 pyproject.toml | 2 +-
 2 files changed, 5 insertions(+), 1 deletion(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 5d3688b64..1e8772fe6 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,5 +1,9 @@
 # Changelog
 
+## 6.20.13 - Dec 27, 2024
+
+* Sanitize phone number for US people scrape.
+
 ## 6.20.12 - Nov 22, 2024
 
 * Use transformers to trim incoming strings at import that are too long for DB columns:
diff --git a/pyproject.toml b/pyproject.toml
index d843ac486..fcc15d276 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "openstates"
-version = "6.20.12"
+version = "6.20.13"
 description = "core infrastructure for the openstates project"
 authors = ["James Turk <dev@jamesturk.net>"]
 license = "MIT"