From 490f0f8e15a3f3f1c4281bf1467a3a347d4e04cd Mon Sep 17 00:00:00 2001
From: Andy Kuny <andrew.kuny@gsa.gov>
Date: Tue, 16 Jan 2024 10:40:01 -0500
Subject: [PATCH] Refine generate_filename method and test

---
 nad_ch/domain/entities.py     | 11 +++++++++--
 nad_ch/use_cases.py           |  2 +-
 tests/domain/test_entities.py | 12 ++++++++++++
 3 files changed, 22 insertions(+), 3 deletions(-)
 create mode 100644 tests/domain/test_entities.py

diff --git a/nad_ch/domain/entities.py b/nad_ch/domain/entities.py
index 0137374..48f77a6 100644
--- a/nad_ch/domain/entities.py
+++ b/nad_ch/domain/entities.py
@@ -1,4 +1,6 @@
 import datetime
+import os
+import re
 
 
 class Entity:
@@ -41,7 +43,12 @@ def __repr__(self):
                 (created: {self.created_at}; updated: {self.updated_at})"
 
     @staticmethod
-    def generate_filename(provider: DataProvider) -> str:
+    def generate_filename(file_path: str, provider: DataProvider) -> str:
+        s = re.sub(r'\W+', '_', provider.name)
+        s = s.lower()
+        s = s.strip('_')
+        formatted_provider_name = re.sub(r'_+', '_', s)
         datetime_str = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
-        filename = f"{provider.name}_{datetime_str}.zip"
+        _, file_extension = os.path.splitext(file_path)
+        filename = f"{formatted_provider_name}_{datetime_str}{file_extension}"
         return filename
diff --git a/nad_ch/use_cases.py b/nad_ch/use_cases.py
index 23d6be7..56ff854 100644
--- a/nad_ch/use_cases.py
+++ b/nad_ch/use_cases.py
@@ -46,7 +46,7 @@ def ingest_data_submission(
         return
 
     try:
-        filename = DataSubmission.generate_filename(provider)
+        filename = DataSubmission.generate_filename(file_path, provider)
         ctx.storage.upload(file_path, filename)
 
         submission = DataSubmission(filename, provider)
diff --git a/tests/domain/test_entities.py b/tests/domain/test_entities.py
new file mode 100644
index 0000000..0e530c8
--- /dev/null
+++ b/tests/domain/test_entities.py
@@ -0,0 +1,12 @@
+import datetime
+from nad_ch.domain.entities import DataProvider, DataSubmission
+
+
+def test_data_submission_generates_filename():
+    provider = DataProvider("Some Provider")
+    filename = DataSubmission.generate_filename("someupload.zip", provider)
+    todays_date = datetime.datetime.now().strftime("%Y%m%d")
+    print(filename)
+    assert filename.startswith("some_provider")
+    assert todays_date in filename
+    assert filename.endswith(".zip")