Many licenses not available in pypi metadata

[kenmalton]

• Package Name: azure-batch (but affecting 149 packages in total)
• Package Version: 15.1.0
• Operating System: n/a
• Python Version: n/a

Describe the bug
The license is not available for many packages in the pypi metadata. I originally found this for azure-batch but have checked that it applies to 149 packages in total.

The License-Expression field has somehow been shoved into the Author Email (see screenshot)

To Reproduce
Steps to reproduce the behavior:

1. Visit azure-batch on pypi and look at the metadata down the left-hand side
2. Look at the json feed to see the License-Expression: MIT tacked onto the end of the author email.

Expected behavior
Correct license and author email metadata on pypi.

Screenshots
If applicable, add screenshots to help explain your problem.

Additional context

This problem affects 149 packages:

azure-ai-agentserver-core, azure-ai-agentserver-ghcopilot, azure-ai-agentserver-invocations, azure-ai-agentserver-optimization, azure-ai-agentserver-responses, azure-ai-contentunderstanding, azure-ai-discovery, azure-ai-language-conversations-authoring, azure-ai-language-questionanswering-authoring, azure-ai-ml, azure-ai-projects, azure-ai-textanalytics-authoring, azure-ai-transcription, azure-ai-translation-text, azure-ai-voicelive, azure-batch, azure-confidentialledger-certificate, azure-core, azure-core-tracing-opentelemetry, azure-eventhub, azure-identity, azure-keyvault-keys, azure-mgmt-advisor, azure-mgmt-agricultureplatform, azure-mgmt-alertsmanagement, azure-mgmt-appcontainers, azure-mgmt-appnetwork, azure-mgmt-artifactsigning, azure-mgmt-attestation, azure-mgmt-automation, azure-mgmt-avs, azure-mgmt-azurearcdata, azure-mgmt-azurestack, azure-mgmt-azurestackhci, azure-mgmt-baremetalinfrastructure, azure-mgmt-batch, azure-mgmt-billingbenefits, azure-mgmt-certificateregistration, azure-mgmt-changeanalysis, azure-mgmt-cloudhealth, azure-mgmt-cognitiveservices, azure-mgmt-commerce, azure-mgmt-communication, azure-mgmt-compute, azure-mgmt-computebulkactions, azure-mgmt-computelimit, azure-mgmt-confidentialledger, azure-mgmt-connectedcache, azure-mgmt-containerregistry, azure-mgmt-containerregistrytasks, azure-mgmt-containerservice, azure-mgmt-cosmosdb, azure-mgmt-customproviders, azure-mgmt-dashboard, azure-mgmt-datafactory, azure-mgmt-dataprotection, azure-mgmt-datashare, azure-mgmt-dellstorage, azure-mgmt-devhub, azure-mgmt-deviceregistry, azure-mgmt-devtestlabs, azure-mgmt-disconnectedoperations, azure-mgmt-discovery, azure-mgmt-domainregistration, azure-mgmt-durabletask, azure-mgmt-edgeactions, azure-mgmt-education, azure-mgmt-elastic, azure-mgmt-elasticsan, azure-mgmt-fileshares, azure-mgmt-hanaonazure, azure-mgmt-hdinsight, azure-mgmt-healthbot, azure-mgmt-horizondb, azure-mgmt-hybridkubernetes, azure-mgmt-imagebuilder, azure-mgmt-iotcentral, azure-mgmt-iotoperations, azure-mgmt-keyvault, azure-mgmt-kubernetesconfiguration-extensions, azure-mgmt-kubernetesconfiguration-extensiontypes, azure-mgmt-kubernetesconfiguration-fluxconfigurations, azure-mgmt-kubernetesconfiguration-privatelinkscopes, azure-mgmt-loganalytics, azure-mgmt-machinelearningservices, azure-mgmt-managedops, azure-mgmt-managedservices, azure-mgmt-managementgroups, azure-mgmt-managementpartner, azure-mgmt-marketplaceordering, azure-mgmt-mongocluster, azure-mgmt-monitorslis, azure-mgmt-monitorworkspaces, azure-mgmt-mysqlflexibleservers, azure-mgmt-netapp, azure-mgmt-network, azure-mgmt-networkcloud, azure-mgmt-networkfunction, azure-mgmt-nginx, azure-mgmt-operationsmanagement, azure-mgmt-oracledatabase, azure-mgmt-paloaltonetworksngfw, azure-mgmt-peering, azure-mgmt-planetarycomputer, azure-mgmt-policyinsights, azure-mgmt-portal, azure-mgmt-postgresqlflexibleservers, azure-mgmt-powerbidedicated, azure-mgmt-purview, azure-mgmt-quota, azure-mgmt-rdbms, azure-mgmt-recoveryservices, azure-mgmt-recoveryservicesbackup, azure-mgmt-recoveryservicesbackup-passivestamp, azure-mgmt-redhatopenshift, azure-mgmt-redisenterprise, azure-mgmt-relationships, azure-mgmt-relay, azure-mgmt-resource, azure-mgmt-resource-changes, azure-mgmt-resource-databoundaries, azure-mgmt-resource-deployments, azure-mgmt-resource-deploymentstacks, azure-mgmt-resource-features, azure-mgmt-resource-links, azure-mgmt-resource-locks, azure-mgmt-resource-managedapplications, azure-mgmt-resource-policy, azure-mgmt-resource-privatelinks, azure-mgmt-resource-subscriptions, azure-mgmt-resourcegraph, azure-mgmt-resourcehealth, azure-mgmt-serialconsole, azure-mgmt-servicefabricmanagedclusters, azure-mgmt-servicegroups, azure-mgmt-sitemanager, azure-mgmt-standbypool, azure-mgmt-storage, azure-mgmt-storagecache, azure-mgmt-storagediscovery, azure-mgmt-storagemover, azure-mgmt-storagesync, azure-mgmt-web, azure-planetarycomputer, azure-postgresql-auth, azure-search-documents, azure-servicebus, azure-storage-extensions, azure-template

[kashifkhan]

Thank you for the feedback @kenmalton, on a quick review seems like the package data is proper. So the culprit might be somewhere else on our side, but I need to confirm that 100% and dig around a bit more.

here is the script that I used to do a quick spot check

import io
import json
import zipfile
import urllib.request
from email.parser import BytesHeaderParser

# get info from pypi API
info = json.load(urllib.request.urlopen('https://pypi.org/pypi/azure-batch/15.1.0/json'))['info']
print('Info from pypi')
print(f"author email: {info['author_email']}")
print(f'license: {info["license"]}')
print(f"license expression: {info['license_expression']}", end="\n\n")


# get info from actual wheel
url = next(u['url'] for u in json.load(urllib.request.urlopen(f'https://pypi.org/pypi/azure-batch/15.1.0/json'))['urls'] if u['packagetype'] == 'bdist_wheel')
whl = zipfile.ZipFile(io.BytesIO(urllib.request.urlopen(url).read()))
meta_path = next(n for n in whl.namelist() if n.endswith('.dist-info/METADATA'))
meta = BytesHeaderParser().parsebytes(whl.read(meta_path))


print('info from wheel')
print(f'Metadata-Version  : {meta["Metadata-Version"]}')
print(f'Author-email      : {meta["Author-email"]}')
print(f'License-Expression:  {meta["License-Expression"]}')

[kashifkhan]

Hi @kenmalton, wanted to provide you with an update: we have issued a fix on our release process to properly pass in the metadata fields. As new versions of the packages are released they should have everything in the right place

[github-actions]

Hi @kenmalton. Thank you for opening this issue and giving us the opportunity to assist. We believe that this has been addressed. If you feel that further discussion is needed, please add a comment with the text "/unresolve" to remove the "issue-addressed" label and continue the conversation.