Update 3rd party license validation (#9450)

* Update 3rd party license * account for different copyright for different versions Co-authored-by: Ofek Lev <ofekmeister@gmail.com>
DataDog · Jun 1, 2021 · 9c2e49d · 9c2e49d
1 parent 6c70fda
commit 9c2e49d
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 4 deletions.
diff --git a/LICENSE-3rdparty.csv b/LICENSE-3rdparty.csv
@@ -19,8 +19,8 @@ cachetools,PyPI,MIT,Thomas Kemmer
 clickhouse-cityhash,PyPI,MIT,Alexander [Amper] Marshalov
 clickhouse-driver,PyPI,MIT,Konstantin Lebedev
 contextlib2,PyPI,PSF,Nick Coghlan
-cryptography,PyPI,Apache-2.0,The cryptography developers
-cryptography,PyPI,BSD-3-Clause,The cryptography developers
+cryptography,PyPI,Apache-2.0,The Python Cryptographic Authority and individual contributors | The cryptography developers
+cryptography,PyPI,BSD-3-Clause,The Python Cryptographic Authority and individual contributors | The cryptography developers
 cx-Oracle,PyPI,BSD-3-Clause,Anthony Tuininga
 ddtrace,PyPI,BSD-3-Clause,"Datadog, Inc."
 dnspython,PyPI,ISC,Bob Halley

diff --git a/datadog_checks_dev/datadog_checks/dev/tooling/commands/validate/licenses.py b/datadog_checks_dev/datadog_checks/dev/tooling/commands/validate/licenses.py
@@ -104,6 +104,7 @@
 
 
 def format_attribution_line(package_name, license_id, package_copyright):
+    package_copyright = ' | '.join(sorted(package_copyright))
     if ',' in package_copyright:
         package_copyright = f'"{package_copyright}"'
 
@@ -138,13 +139,13 @@ async def get_data(url):
 
 
 async def scrape_license_data(urls):
-    package_data = defaultdict(lambda: {'copyright': '', 'licenses': [], 'classifiers': set()})
+    package_data = defaultdict(lambda: {'copyright': set(), 'licenses': [], 'classifiers': set()})
 
     async with Pool() as pool:
         async for package_name, package_copyright, package_license, license_classifiers in pool.map(get_data, urls):
             data = package_data[package_name]
             if package_copyright:
-                data['copyright'] = package_copyright
+                data['copyright'].add(package_copyright)
 
             data['classifiers'].update(license_classifiers)
             if package_license: