Append "cwltool" to HTTP User-Agent string (#1977)

Also appends the program name, if not "cwltool"

Should work for calrissian, reana-cwl-runner, and arvados-cwl-runner, but not
toil-cwl-runner which doesn't call the cwltool.main.main() function.

Probably works for streamflow, but I haven't tested that.

---------

Co-authored-by: Michael R. Crusoe <[email protected]>

Author: svonworl

cwltool/main.py

@@ -33,6 +33,7 @@
 
 import argcomplete
 import coloredlogs
+import requests
 import ruamel.yaml
 from ruamel.yaml.comments import CommentedMap, CommentedSeq
 from ruamel.yaml.main import YAML
@@ -173,6 +174,14 @@ def _signal_handler(signum: int, _: Any) -> None:
     sys.exit(signum)
 
 
+def append_word_to_default_user_agent(word: str) -> None:
+    """Append the specified word to the requests http user agent string if it's not already there."""
+    original_function = requests.utils.default_user_agent
+    suffix = f" {word}"
+    if not original_function().endswith(suffix):
+        requests.utils.default_user_agent = lambda *args: original_function(*args) + suffix
+
+
 def generate_example_input(
     inptype: Optional[CWLOutputType],
     default: Optional[CWLOutputType],
@@ -979,6 +988,12 @@ def main(
     workflowobj = None
     prov_log_handler: Optional[logging.StreamHandler[ProvOut]] = None
     global docker_exe
+
+    user_agent = "cwltool"
+    if user_agent not in (progname := os.path.basename(sys.argv[0])):
+        user_agent += f" {progname}"  # append the real program name as well
+    append_word_to_default_user_agent(user_agent)
+
     try:
         if args is None:
             if argsl is None:

tests/test_user_agent.py

@@ -0,0 +1,24 @@
+import requests
+
+from cwltool.main import append_word_to_default_user_agent, main
+
+
+def get_user_agent() -> str:
+    return requests.utils.default_headers()["User-Agent"]
+
+
+def test_cwltool_in_user_agent() -> None:
+    """python-requests HTTP User-Agent should include the string 'cwltool'."""
+    try:
+        assert main(["--version"]) == 0
+    except SystemExit as err:
+        assert err.code == 0
+    assert "cwltool" in get_user_agent()
+
+
+def test_append_word_to_default_user_agent() -> None:
+    """Confirm that append_word_to_default_user_agent works."""
+    word_to_append = "foobar123"
+    assert word_to_append not in get_user_agent()
+    append_word_to_default_user_agent(word_to_append)
+    assert word_to_append in get_user_agent()