opam: Allow defining where to actually install the opam_root folder

Related to T3590

swh/lister/opam/lister.py

@@ -7,7 +7,6 @@ import io
 import logging
 import os
 from subprocess import PIPE, Popen, call
-import tempfile
 from typing import Iterator, Optional
 
 from swh.lister.pattern import StatelessLister
@@ -46,6 +45,7 @@ class OpamLister(StatelessLister[PageType]):
         url: str,
         instance: Optional[str] = None,
         credentials: CredentialsType = None,
+        opam_root: str = "/tmp/opam/",
     ):
         super().__init__(
             scheduler=scheduler, credentials=credentials, url=url, instance=instance,
@@ -53,7 +53,7 @@ class OpamLister(StatelessLister[PageType]):
         self.env = os.environ.copy()
         # Opam root folder is initialized in the :meth:`get_pages` method as no
         # side-effect should happen in the constructor to ease instantiation
-        self.opamroot = tempfile.mkdtemp(prefix="swh_opam_lister")
+        self.opamroot = os.path.join(opam_root, self.instance)
 
     def get_pages(self) -> Iterator[PageType]:
         # Initialize the opam root directory with the opam instance data to list.

swh/lister/opam/tests/test_lister.py

@@ -4,6 +4,7 @@
 # See top-level LICENSE file for more information
 
 import io
+from tempfile import mkdtemp
 from unittest.mock import MagicMock
 
 import pytest
@@ -33,8 +34,9 @@ def test_lister_opam_optional_instance(swh_scheduler):
     netloc = "opam.ocaml.org"
     instance_url = f"https://{netloc}"
 
-    lister = OpamLister(swh_scheduler, url=instance_url)
+    lister = OpamLister(swh_scheduler, url=instance_url,)
     assert lister.instance == netloc
+    assert lister.opamroot.endswith(lister.instance)
 
 
 def test_urls(swh_scheduler, mock_opam):
@@ -42,7 +44,12 @@ def test_urls(swh_scheduler, mock_opam):
 
     instance_url = "https://opam.ocaml.org"
 
-    lister = OpamLister(swh_scheduler, url=instance_url, instance="opam")
+    lister = OpamLister(
+        swh_scheduler,
+        url=instance_url,
+        instance="opam",
+        opam_root=mkdtemp(prefix="swh_opam_lister"),
+    )
     assert lister.instance == "opam"
 
     # call the lister and get all listed origins urls
@@ -70,7 +77,12 @@ def test_urls(swh_scheduler, mock_opam):
 def test_opam_binary(datadir, swh_scheduler):
     instance_url = f"file://{datadir}/fake_opam_repo"
 
-    lister = OpamLister(swh_scheduler, url=instance_url, instance="fake")
+    lister = OpamLister(
+        swh_scheduler,
+        url=instance_url,
+        instance="fake",
+        opam_root=mkdtemp(prefix="swh_opam_lister"),
+    )
 
     stats = lister.run()