Fix Grobid/TEI generation (#669)

* update

* test

* notify on grobid version mismatch

* test

* update

* show running version

* update

* update

* reset

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* revisions

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* add note/explanation

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Author: geritwagner

colrev/env/grobid_service.py

@@ -15,7 +15,9 @@ class GrobidService:
     """An environment service for machine readability/annotation (PDF to TEI conversion)"""
 
     GROBID_URL = "http://localhost:8070"
-    GROBID_IMAGE = "lfoppiano/grobid:latest-crf"
+    # Important: do not use :latest versions or :SNAPSHOT versions
+    # as they may change without notice
+    GROBID_IMAGE = "lfoppiano/grobid:0.8.2"
 
     def __init__(self) -> None:
         colrev.env.docker_manager.DockerManager.build_docker_image(
@@ -24,6 +26,17 @@ def __init__(self) -> None:
         self.start()
         self.check_grobid_availability()
 
+    def _ensure_correct_version(self) -> None:
+        response = requests.get(self.GROBID_URL + "/api/version", timeout=10)
+        running_version = response.json()["version"]
+        if running_version != self.GROBID_IMAGE.split(":")[1]:
+            logging.warning(
+                "GROBID version mismatch. Expected: %s, currently running: %s",
+                self.GROBID_IMAGE.split(":")[1],
+                running_version,
+            )
+            raise Exception
+
     def check_grobid_availability(self, *, wait: bool = True) -> bool:
         """Check whether the GROBID service is available"""
         i = 0
@@ -33,6 +46,9 @@ def check_grobid_availability(self, *, wait: bool = True) -> bool:
             try:
                 ret = requests.get(self.GROBID_URL + "/api/isalive", timeout=30)
                 if ret.text == "true":
+                    # When GROBID is running, it may not be the same version as expected
+                    # in self.GROBID_IMAGE, possibly leading to failing tests.
+                    self._ensure_correct_version()
                     return True
             except requests.exceptions.ConnectionError:
                 pass

tests/1_env/tei_test.py

@@ -44,27 +44,36 @@ def test_tei_creation(script_loc) -> None:  # type: ignore
     tei_file = script_loc.parent.joinpath("data/WagnerLukyanenkoParEtAl2022.tei.xml")
     pdf_path = script_loc.parent.joinpath("data/WagnerLukyanenkoParEtAl2022.pdf")
 
-    tei_file.unlink(missing_ok=True)
+    tmp_tei_file = tei_file.with_name(tei_file.stem + "_tmp.tei.xml")
+    if tmp_tei_file.exists():
+        tmp_tei_file.unlink(missing_ok=True)
+    tei_file.rename(tmp_tei_file)
 
-    colrev.env.tei_parser.TEIParser(pdf_path=pdf_path, tei_path=tei_file)
+    try:
 
-    with open(tei_file) as file:
-        tei_content = file.read()
+        colrev.env.tei_parser.TEIParser(pdf_path=pdf_path, tei_path=tei_file)
 
-    tei_content = re.sub(
-        r'(ident="GROBID" when=")[^"]+(">)', r"\g<1>NA\g<2>", tei_content
-    )
+        with open(tei_file) as file:
+            tei_content = file.read()
+
+        tei_content = re.sub(
+            r'(ident="GROBID" when=")[^"]+(">)', r"\g<1>NA\g<2>", tei_content
+        )
 
-    with open(tei_file, "w") as file:
-        file.write(tei_content)
+        with open(tei_file, "w") as file:
+            file.write(tei_content)
+    except Exception as exc:
+        print("Restoring original TEI file")
+        tmp_tei_file.rename(tei_file)
+        raise exc
 
 
 @pytest.mark.skipif(
     platform.system() != "Linux", reason="Docker tests only run on Linux runners"
 )
 def test_tei_version(tei_doc) -> None:  # type: ignore
     """Test the tei version"""
-    assert "0.8.3-SNAPSHOT" == tei_doc.get_grobid_version()
+    assert "0.8.2" == tei_doc.get_grobid_version()
 
 
 @pytest.mark.skipif(