opensafely 2.0.4__tar.gz → 2.1.1__tar.gz

{opensafely-2.0.4 → opensafely-2.1.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: opensafely
-Version: 2.0.4
+Version: 2.1.1
 Summary: Command line tool for running OpenSAFELY studies locally.
 Author-email: OpenSAFELY <tech@opensafely.org>
 License: GPL-3.0

opensafely-2.1.1/opensafely/VERSION

@@ -0,0 +1 @@
+v2.1.1

{opensafely-2.0.4 → opensafely-2.1.1}/opensafely/_vendor/pipeline/models.py

@@ -212,8 +212,11 @@ class Action:
 
         if re.match(r"cohortextractor:\S+ generate_cohort", run.raw):
             validate_cohortextractor_outputs(action_id, action)
-        if re.match(r"(ehrql|databuilder):\S+ generate[-_]dataset", run.raw):
-            validate_ehrql_outputs(action_id, action)
+        if match := re.match(
+            r"(ehrql|databuilder):\S+ generate[-_](?P<ehrql_type>dataset|measures)",
+            run.raw,
+        ):
+            validate_ehrql_outputs(action_id, action, match.group("ehrql_type"))
 
         return action
 

{opensafely-2.0.4 → opensafely-2.1.1}/opensafely/_vendor/pipeline/validation.py

@@ -156,14 +156,16 @@ def validate_cohortextractor_outputs(action_id: str, action: Action) -> None:
         )
 
 
-def validate_ehrql_outputs(action_id: str, action: Action) -> None:
+def validate_ehrql_outputs(action_id: str, action: Action, ehrql_type: str) -> None:
     """
     Check ehrQL's output config is valid for this command
 
     We can't validate outputs in the Action or Outputs models because we need
     to look up other fields (eg run).
     """
-    if action.outputs.moderately_sensitive or action.outputs.minimally_sensitive:
+    if ehrql_type == "dataset" and (
+        action.outputs.moderately_sensitive or action.outputs.minimally_sensitive
+    ):
         raise ValidationError(
             f"`{action_id}` action uses `generate-dataset` and so all outputs must "
             f"be labelled `highly_sensitive`"
@@ -173,11 +175,18 @@ def validate_ehrql_outputs(action_id: str, action: Action) -> None:
     if not output_spec:
         raise ValidationError(
             f"`{action_id}` action does not provide an `--output` argument specifying "
-            f"where the results of `generate-dataset` should be stored"
+            f"where the results of `generate-{ehrql_type}` should be stored"
         )
 
-    output_patterns = (action.outputs.highly_sensitive or {}).values()
-    if not output_patterns_match_spec(output_spec, list(output_patterns)):
+    if ehrql_type == "dataset":
+        output_patterns = list((action.outputs.highly_sensitive or {}).values())
+    else:
+        assert ehrql_type == "measures"
+        output_patterns = list((action.outputs.highly_sensitive or {}).values()) + list(
+            (action.outputs.moderately_sensitive or {}).values()
+        )
+
+    if not output_patterns_match_spec(output_spec, output_patterns):
         expected = expected_pattern_for_spec(output_spec)
         formatted_patterns = "                   \n".join(output_patterns)
         raise ValidationError(

{opensafely-2.0.4 → opensafely-2.1.1}/opensafely/jobrunner/actions.py

@@ -4,9 +4,6 @@ import shlex
 
 from opensafely._vendor.pipeline.exceptions import ProjectValidationError
 from opensafely._vendor.pipeline.models import Action
-from opensafely._vendor.pipeline.outputs import get_output_dirs
-
-from opensafely.jobrunner.lib.path_utils import ensure_unix_path
 
 
 class UnknownActionError(ProjectValidationError):
@@ -54,27 +51,6 @@ def get_action_specification(config, action_id, using_dummy_data_backend=False):
             json.dumps(action_spec.config).replace("'", r"\u0027"),
         ]
 
-    # Special case handling for the `cohortextractor generate_cohort` command
-    if is_cohortextractor_generate_cohort(run_parts):
-        # Set the size of the dummy data population, if that's what we're
-        # generating.  Possibly this should be moved to the study definition
-        # anyway, which would make this unnecessary.
-        if using_dummy_data_backend:
-            if action_spec.dummy_data_file is not None:
-                run_parts.append(
-                    f"--dummy-data-file={ensure_unix_path(action_spec.dummy_data_file)}"
-                )
-            else:
-                size = config.expectations.population_size
-                run_parts.append(f"--expectations-population={size}")
-
-        output_dirs = get_output_dirs(action_spec.outputs)
-
-        if len(output_dirs) == 1:
-            # Automatically configure the cohortextractor to produce output in the
-            # directory the `outputs` spec is expecting.
-            run_parts.append(f"--output-dir={output_dirs[0]}")
-
     run_command = shlex.join(run_parts)
 
     return ActionSpecification(
@@ -83,11 +59,3 @@ def get_action_specification(config, action_id, using_dummy_data_backend=False):
         outputs=action_spec.outputs.dict(),
         action=action_spec,
     )
-
-
-def is_cohortextractor_generate_cohort(args):
-    return (
-        len(args) > 1
-        and args[0].startswith("cohortextractor:")
-        and args[1] == "generate_cohort"
-    )

{opensafely-2.0.4 → opensafely-2.1.1}/opensafely/jobrunner/config.py

@@ -76,7 +76,6 @@ else:
     USING_DUMMY_DATA_BACKEND = BACKEND == "expectations"
 
 ALLOWED_IMAGES = {
-    "cohortextractor",
     "databuilder",
     "ehrql",
     "stata-mp",
@@ -214,15 +213,36 @@ ENABLE_MAINTENANCE_MODE_THREAD = os.environ.get(
 
 
 # Map known exit codes to user-friendly messages
+# Taken from https://github.com/opensafely-core/job-runner/blob/dd0c916db15168d3d0be0134c1dc1fbb7f54785a/controller/config.py#L160
 DATABASE_EXIT_CODES = {
-    # Custom database-related exit codes return from cohortextractor, see
-    # https://github.com/opensafely-core/cohort-extractor/blob/0a314a909817dbcc48907643e0b6eeff319337db/cohortextractor/cohortextractor.py#L787
+    # Custom database-related exit codes return from ehrQL, see e.g.
+    # https://github.com/opensafely-core/ehrql/blob/868e07b9d3638996c14a573030ca946765d8fdf9/ehrql/backends/tpp.py#L229-L254
+    # These are unlikely to be hit since the CLI always uses dummy data and ehrql's
+    # in-memory engine
     3: (
         "A transient database error occurred, your job may run "
         "if you try it again, if it keeps failing then contact tech support"
     ),
     4: "New data is being imported into the database, please try again in a few hours",
     5: "Something went wrong with the database, please contact tech support",
+    6: (
+        "Your ehrQL is valid but it produced SQL which is too complex for the database"
+        " to run; please contact tech-support so we can help you simplify it."
+    ),
+    # Other ehrQL exit codes with specific meanings, see:
+    # https://github.com/opensafely-core/ehrql/blob/868e07b9d3638996c14a573030ca946765d8fdf9/ehrql/__main__.py#L112
+    # https://github.com/opensafely-core/ehrql/blob/868e07b9d3638996c14a573030ca946765d8fdf9/ehrql/exceptions.py#L66
+    # Messages have been modified to remove instructions that only apply to real
+    # jobs run in the backend (e.g. to check code runs locally, or to contact tech-support)
+    2: ("There was a problem with the command line arguments in your ehrQL action"),
+    10: "There was a problem reading your ehrQL code",
+    11: "There was a problem reading one of the supplied data files",
+    12: "You do not have the required permissions for the ehrQL you are trying to run",
+    13: "Assurance tests failed",
+    14: (
+        "Your job was making such slow progress that it was unlikely to complete "
+        "successfully"
+    ),
 }
 
 

{opensafely-2.0.4 → opensafely-2.1.1}/opensafely/jobrunner/models.py

@@ -212,7 +212,7 @@ class Job:
     commit: str = None
     # Name of workspace (effectively, the output directory)
     workspace: str = None
-    # Only applicable to "generate_cohort" jobs: the name of the database to
+    # Only applicable to "generate-dataset" jobs: the name of the database to
     # query against
     database_name: str = None
     # Name of the action (one of the keys in the `actions` dict in

{opensafely-2.0.4 → opensafely-2.1.1}/opensafely/pull.py

@@ -170,8 +170,6 @@ def remove_deprecated_images(local_images):
 def get_default_version_for_image(name):
     if name in ["python", "r", "rstudio"]:
         return "v2"
-    elif name == "cohortextractor":
-        return "latest"
     else:
         return "v1"
 

{opensafely-2.0.4 → opensafely-2.1.1}/opensafely.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: opensafely
-Version: 2.0.4
+Version: 2.1.1
 Summary: Command line tool for running OpenSAFELY studies locally.
 Author-email: OpenSAFELY <tech@opensafely.org>
 License: GPL-3.0

{opensafely-2.0.4 → opensafely-2.1.1}/tests/test_pull.py

@@ -44,7 +44,6 @@ def test_default_no_local_images(run, capsys):
 def test_default_no_local_images_force(run, capsys):
     run.expect(["docker", "info"])
     expect_local_images(run)
-    run.expect(["docker", "pull", tag("cohortextractor", version="latest")])
     run.expect(["docker", "pull", tag("ehrql")])
     run.expect(["docker", "pull", tag("jupyter")])
     run.expect(["docker", "pull", tag("python", version="v2")])
@@ -67,7 +66,6 @@ def test_default_no_local_images_force(run, capsys):
     out, err = capsys.readouterr()
     assert err == ""
     assert out.splitlines() == [
-        "Updating OpenSAFELY cohortextractor:latest image",
         "Updating OpenSAFELY ehrql:v1 image",
         "Updating OpenSAFELY jupyter:v1 image",
         "Updating OpenSAFELY python:v2 image",

opensafely-2.0.4/opensafely/VERSION

@@ -1 +0,0 @@
-v2.0.4