metaflow 2.15.0__py2.py3-none-any.whl → 2.15.1__py2.py3-none-any.whl

metaflow/__init__.py

@@ -119,6 +119,13 @@ from .includefile import IncludeFile
 # Decorators
 from .decorators import step, _import_plugin_decorators
 
+# Config parsers
+from .plugins.pypi.parsers import (
+    requirements_txt_parser,
+    pyproject_toml_parser,
+    conda_environment_yml_parser,
+)
+
 # this auto-generates decorator functions from Decorator objects
 # in the top-level metaflow namespace
 _import_plugin_decorators(globals())

metaflow/cmd/code/__init__.py

@@ -0,0 +1,230 @@
+import os
+import shutil
+import sys
+from subprocess import PIPE, CompletedProcess, run
+from tempfile import TemporaryDirectory
+from typing import Any, Callable, List, Mapping, Optional, cast
+
+from metaflow import Run
+from metaflow._vendor import click
+from metaflow.cli import echo_always
+
+
+@click.group()
+def cli():
+    pass
+
+
+@cli.group(help="Access, compare, and manage code associated with Metaflow runs.")
+def code():
+    pass
+
+
+def echo(line: str) -> None:
+    echo_always(line, err=True, fg="magenta")
+
+
+def extract_code_package(runspec: str) -> TemporaryDirectory:
+    try:
+        mf_run = Run(runspec, _namespace_check=False)
+        echo(f"✅  Run *{runspec}* found, downloading code..")
+    except Exception as e:
+        echo(f"❌  Run **{runspec}** not found")
+        raise e
+
+    if mf_run.code is None:
+        echo(
+            f"❌  Run **{runspec}** doesn't have a code package. Maybe it's a local run?"
+        )
+        raise RuntimeError("no code package found")
+
+    return mf_run.code.extract()
+
+
+def perform_diff(
+    source_dir: str,
+    target_dir: Optional[str] = None,
+    output: bool = False,
+    **kwargs: Mapping[str, Any],
+) -> Optional[List[str]]:
+    if target_dir is None:
+        target_dir = os.getcwd()
+
+    diffs = []
+    for dirpath, dirnames, filenames in os.walk(source_dir, followlinks=True):
+        for fname in filenames:
+            # NOTE: the paths below need to be set up carefully
+            # for the `patch` command to work. Better not to touch
+            # the directories below. If you must, test that patches
+            # work after your changes.
+            #
+            # target_file is the git repo in the current working directory
+            rel = os.path.relpath(dirpath, source_dir)
+            target_file = os.path.join(rel, fname)
+            # source_file is the run file loaded in a tmp directory
+            source_file = os.path.join(dirpath, fname)
+
+            if sys.stdout.isatty() and not output:
+                color = ["--color"]
+            else:
+                color = ["--no-color"]
+
+            if os.path.exists(os.path.join(target_dir, target_file)):
+                cmd = (
+                    ["git", "diff", "--no-index", "--exit-code"]
+                    + color
+                    + [
+                        target_file,
+                        source_file,
+                    ]
+                )
+                result: CompletedProcess = run(
+                    cmd, text=True, stdout=PIPE, cwd=target_dir
+                )
+                if result.returncode == 0:
+                    if not output:
+                        echo(f"✅  {target_file} is identical, skipping")
+                    continue
+
+                if output:
+                    diffs.append(result.stdout)
+                else:
+                    run(["less", "-R"], input=result.stdout, text=True)
+            else:
+                if not output:
+                    echo(f"❗  {target_file} not in the target directory, skipping")
+    return diffs if output else None
+
+
+def run_op(
+    runspec: str, op: Callable[..., Optional[List[str]]], **op_args: Mapping[str, Any]
+) -> Optional[List[str]]:
+    tmp = None
+    try:
+        tmp = extract_code_package(runspec)
+        return op(tmp.name, **op_args)
+    finally:
+        if tmp and os.path.exists(tmp.name):
+            shutil.rmtree(tmp.name)
+
+
+def run_op_diff_runs(
+    source_run_pathspec: str, target_run_pathspec: str, **op_args: Mapping[str, Any]
+) -> Optional[List[str]]:
+    source_tmp = None
+    target_tmp = None
+    try:
+        source_tmp = extract_code_package(source_run_pathspec)
+        target_tmp = extract_code_package(target_run_pathspec)
+        return perform_diff(source_tmp.name, target_tmp.name, **op_args)
+    finally:
+        for d in [source_tmp, target_tmp]:
+            if d and os.path.exists(d.name):
+                shutil.rmtree(d.name)
+
+
+def op_diff(tmpdir: str, **kwargs: Mapping[str, Any]) -> Optional[List[str]]:
+    kwargs_dict = dict(kwargs)
+    target_dir = cast(Optional[str], kwargs_dict.pop("target_dir", None))
+    output: bool = bool(kwargs_dict.pop("output", False))
+    op_args: Mapping[str, Any] = {**kwargs_dict}
+    return perform_diff(tmpdir, target_dir=target_dir, output=output, **op_args)
+
+
+def op_pull(tmpdir: str, dst: str, **op_args: Mapping[str, Any]) -> None:
+    if os.path.exists(dst):
+        echo(f"❌  Directory *{dst}* already exists")
+    else:
+        shutil.move(tmpdir, dst)
+        echo(f"Code downloaded to *{dst}*")
+
+
+def op_patch(tmpdir: str, dst: str, **kwargs: Mapping[str, Any]) -> None:
+    diffs = perform_diff(tmpdir, output=True) or []
+    with open(dst, "w", encoding="utf-8") as f:
+        for out in diffs:
+            out = out.replace(tmpdir, "/.")
+            out = out.replace("+++ b/./", "+++ b/")
+            out = out.replace("--- b/./", "--- b/")
+            out = out.replace("--- a/./", "--- a/")
+            out = out.replace("+++ a/./", "+++ a/")
+            f.write(out)
+    echo(f"Patch saved in *{dst}*")
+    path = run(
+        ["git", "rev-parse", "--show-prefix"], text=True, stdout=PIPE
+    ).stdout.strip()
+    if path:
+        diropt = f" --directory={path.rstrip('/')}"
+    else:
+        diropt = ""
+    echo("Apply the patch by running:")
+    echo_always(
+        f"git apply --verbose{diropt} {dst}", highlight=True, bold=True, err=True
+    )
+
+
+@code.command()
+@click.argument("run_pathspec")
+def diff(run_pathspec: str, **kwargs: Mapping[str, Any]) -> None:
+    """
+    Do a 'git diff' of the current directory and a Metaflow run.
+    """
+    _ = run_op(run_pathspec, op_diff, **kwargs)
+
+
+@code.command()
+@click.argument("source_run_pathspec")
+@click.argument("target_run_pathspec")
+def diff_runs(
+    source_run_pathspec: str, target_run_pathspec: str, **kwargs: Mapping[str, Any]
+) -> None:
+    """
+    Do a 'git diff' between two Metaflow runs.
+    """
+    _ = run_op_diff_runs(source_run_pathspec, target_run_pathspec, **kwargs)
+
+
+@code.command()
+@click.argument("run_pathspec")
+@click.option(
+    "--dir", help="Destination directory (default: {run_pathspec}_code)", default=None
+)
+def pull(
+    run_pathspec: str, dir: Optional[str] = None, **kwargs: Mapping[str, Any]
+) -> None:
+    """
+    Pull the code of a Metaflow run.
+    """
+    if dir is None:
+        dir = run_pathspec.lower().replace("/", "_") + "_code"
+    op_args: Mapping[str, Any] = {**kwargs, "dst": dir}
+    run_op(run_pathspec, op_pull, **op_args)
+
+
+@code.command()
+@click.argument("run_pathspec")
+@click.option(
+    "--file_path",
+    help="Patch file name. If not provided, defaults to a sanitized version of RUN_PATHSPEC "
+    "with slashes replaced by underscores, plus '.patch'.",
+    show_default=False,
+)
+@click.option(
+    "--overwrite", is_flag=True, help="Overwrite the patch file if it exists."
+)
+def patch(
+    run_pathspec: str,
+    file_path: Optional[str] = None,
+    overwrite: bool = False,
+    **kwargs: Mapping[str, Any],
+) -> None:
+    """
+    Create a patch by comparing current dir with a Metaflow run.
+    """
+    if file_path is None:
+        file_path = run_pathspec.lower().replace("/", "_") + ".patch"
+    if os.path.exists(file_path) and not overwrite:
+        echo(f"File *{file_path}* already exists. To overwrite, specify --overwrite.")
+        return
+    op_args: Mapping[str, Any] = {**kwargs, "dst": file_path}
+    run_op(run_pathspec, op_patch, **op_args)

metaflow/cmd/develop/stub_generator.py

@@ -1133,13 +1133,16 @@ class StubGenerator:
             result = result[1:]
         # Add doc to first and last overloads. Jedi uses the last one and pycharm
         # the first one. Go figure.
+        result_docstring = docs["func_doc"]
+        if docs["param_doc"]:
+            result_docstring += "\nParameters\n----------\n" + docs["param_doc"]
         result[0] = (
             result[0][0],
-            docs["func_doc"] + "\nParameters\n----------\n" + docs["param_doc"],
+            result_docstring,
         )
         result[-1] = (
             result[-1][0],
-            docs["func_doc"] + "\nParameters\n----------\n" + docs["param_doc"],
+            result_docstring,
         )
         return result
 

metaflow/cmd/main_cli.py

@@ -67,6 +67,7 @@ CMDS_DESC = [
     ("configure", ".configure_cmd.cli"),
     ("tutorials", ".tutorials_cmd.cli"),
     ("develop", ".develop.cli"),
+    ("code", ".code.cli"),
 ]
 
 process_cmds(globals())

metaflow/metaflow_config.py

@@ -381,6 +381,8 @@ KUBERNETES_DISK = from_conf("KUBERNETES_DISK", None)
 # Default kubernetes QoS class
 KUBERNETES_QOS = from_conf("KUBERNETES_QOS", "burstable")
 
+# Architecture of kubernetes nodes - used for @conda/@pypi in metaflow-dev
+KUBERNETES_CONDA_ARCH = from_conf("KUBERNETES_CONDA_ARCH")
 ARGO_WORKFLOWS_KUBERNETES_SECRETS = from_conf("ARGO_WORKFLOWS_KUBERNETES_SECRETS", "")
 ARGO_WORKFLOWS_ENV_VARS_TO_SKIP = from_conf("ARGO_WORKFLOWS_ENV_VARS_TO_SKIP", "")
 

metaflow/metaflow_environment.py

@@ -6,7 +6,7 @@ from .util import get_username
 from . import metaflow_version
 from metaflow.exception import MetaflowException
 from metaflow.extension_support import dump_module_info
-from metaflow.mflog import BASH_MFLOG
+from metaflow.mflog import BASH_MFLOG, BASH_FLUSH_LOGS
 from . import R
 
 
@@ -159,6 +159,7 @@ class MetaflowEnvironment(object):
     def get_package_commands(self, code_package_url, datastore_type):
         cmds = [
             BASH_MFLOG,
+            BASH_FLUSH_LOGS,
             "mflog 'Setting up task environment.'",
             self._get_install_dependencies_cmd(datastore_type),
             "mkdir metaflow",
@@ -176,6 +177,7 @@ class MetaflowEnvironment(object):
             "fi" % code_package_url,
             "TAR_OPTIONS='--warning=no-timestamp' tar xf job.tar",
             "mflog 'Task is starting.'",
+            "flush_mflogs",
         ]
         return cmds
 

metaflow/mflog/__init__.py

@@ -44,6 +44,8 @@ BASH_MFLOG = (
 BASH_SAVE_LOGS_ARGS = ["python", "-m", "metaflow.mflog.save_logs"]
 BASH_SAVE_LOGS = " ".join(BASH_SAVE_LOGS_ARGS)
 
+BASH_FLUSH_LOGS = "flush_mflogs(){ " f"{BASH_SAVE_LOGS}; " "}"
+
 
 # this function returns a bash expression that redirects stdout
 # and stderr of the given bash expression to mflog.tee
@@ -63,7 +65,7 @@ def bash_capture_logs(bash_expr, var_transform=None):
 # update_delay determines how often logs should be uploaded to S3
 # as a function of the task execution time
 
-MIN_UPDATE_DELAY = 1.0  # the most frequent update interval
+MIN_UPDATE_DELAY = 0.25  # the most frequent update interval
 MAX_UPDATE_DELAY = 30.0  # the least frequent update interval
 
 
@@ -110,7 +112,6 @@ def export_mflog_env_vars(
 
 def tail_logs(prefix, stdout_tail, stderr_tail, echo, has_log_updates):
     def _available_logs(tail, stream, echo, should_persist=False):
-        # print the latest batch of lines
         try:
             for line in tail:
                 if should_persist:
@@ -128,7 +129,7 @@ def tail_logs(prefix, stdout_tail, stderr_tail, echo, has_log_updates):
 
     start_time = time.time()
     next_log_update = start_time
-    log_update_delay = 1
+    log_update_delay = update_delay(0)
     while has_log_updates():
         if time.time() > next_log_update:
             _available_logs(stdout_tail, "stdout", echo)

metaflow/plugins/kubernetes/kubernetes_decorator.py

@@ -29,6 +29,7 @@ from metaflow.metaflow_config import (
     KUBERNETES_SHARED_MEMORY,
     KUBERNETES_TOLERATIONS,
     KUBERNETES_QOS,
+    KUBERNETES_CONDA_ARCH,
 )
 from metaflow.plugins.resources_decorator import ResourcesDecorator
 from metaflow.plugins.timeout_decorator import get_run_time_limit_for_task
@@ -158,7 +159,7 @@ class KubernetesDecorator(StepDecorator):
 
     # Conda environment support
     supports_conda_environment = True
-    target_platform = "linux-64"
+    target_platform = KUBERNETES_CONDA_ARCH or "linux-64"
 
     def init(self):
         super(KubernetesDecorator, self).init()

metaflow/plugins/kubernetes/kubernetes_jobsets.py

@@ -319,6 +319,8 @@ class RunningJobSet(object):
     def kill(self):
         plural = "jobsets"
         client = self._client.get()
+        if not (self.is_running or self.is_waiting):
+            return
         try:
             # Killing the control pod will trigger the jobset to mark everything as failed.
             # Since jobsets have a successPolicy set to `All` which ensures that everything has

metaflow/plugins/pypi/bootstrap.py

@@ -8,6 +8,7 @@ import subprocess
 import sys
 import tarfile
 import time
+import platform
 from urllib.error import URLError
 from urllib.request import urlopen
 from metaflow.metaflow_config import DATASTORE_LOCAL_DIR, CONDA_USE_FAST_INIT
@@ -36,29 +37,6 @@ def timer(func):
 
 
 if __name__ == "__main__":
-    # TODO: Detect architecture on the fly when dealing with arm architectures.
-    # ARCH=$(uname -m)
-    # OS=$(uname)
-
-    # if [[ "$OS" == "Linux" ]]; then
-    #     PLATFORM="linux"
-    #     if [[ "$ARCH" == "aarch64" ]]; then
-    #         ARCH="aarch64";
-    #     elif [[ $ARCH == "ppc64le" ]]; then
-    #         ARCH="ppc64le";
-    #     else
-    #         ARCH="64";
-    #     fi
-    # fi
-
-    # if [[ "$OS" == "Darwin" ]]; then
-    #     PLATFORM="osx";
-    #     if [[ "$ARCH" == "arm64" ]]; then
-    #         ARCH="arm64";
-    #     else
-    #         ARCH="64"
-    #     fi
-    # fi
 
     def run_cmd(cmd, stdin_str=None):
         result = subprocess.run(
@@ -350,12 +328,25 @@ if __name__ == "__main__":
         cmd = f"fast-initializer --prefix {prefix} --packages-dir {pkgs_dir}"
         run_cmd(cmd, all_package_urls)
 
-    if len(sys.argv) != 5:
-        print("Usage: bootstrap.py <flow_name> <id> <datastore_type> <architecture>")
+    if len(sys.argv) != 4:
+        print("Usage: bootstrap.py <flow_name> <id> <datastore_type>")
         sys.exit(1)
 
     try:
-        _, flow_name, id_, datastore_type, architecture = sys.argv
+        _, flow_name, id_, datastore_type = sys.argv
+
+        system = platform.system().lower()
+        arch_machine = platform.machine().lower()
+
+        if system == "darwin" and arch_machine == "arm64":
+            architecture = "osx-arm64"
+        elif system == "darwin":
+            architecture = "osx-64"
+        elif system == "linux" and arch_machine == "aarch64":
+            architecture = "linux-aarch64"
+        else:
+            # default fallback
+            architecture = "linux-64"
 
         prefix = os.path.join(os.getcwd(), architecture, id_)
         pkgs_dir = os.path.join(os.getcwd(), ".pkgs")

metaflow/plugins/pypi/conda_environment.py

@@ -190,7 +190,6 @@ class CondaEnvironment(MetaflowEnvironment):
         #  4. Start PyPI solves in parallel after each conda environment is created
         #  5. Download PyPI packages sequentially
         #  6. Create and cache PyPI environments in parallel
-
         with ThreadPoolExecutor() as executor:
             # Start all conda solves in parallel
             conda_futures = [
@@ -213,14 +212,14 @@ class CondaEnvironment(MetaflowEnvironment):
 
                 # Queue PyPI solve to start after conda create
                 if result[0] in pypi_envs:
+                    # solve pypi envs uniquely
+                    pypi_env = pypi_envs.pop(result[0])
 
                     def pypi_solve(env):
                         create_future.result()  # Wait for conda create
                         return solve(*env, "pypi")
 
-                    pypi_futures.append(
-                        executor.submit(pypi_solve, pypi_envs[result[0]])
-                    )
+                    pypi_futures.append(executor.submit(pypi_solve, pypi_env))
 
             # Process PyPI results sequentially for downloads
             for solve_future in pypi_futures:
@@ -242,7 +241,7 @@ class CondaEnvironment(MetaflowEnvironment):
         if id_:
             # bootstrap.py is responsible for ensuring the validity of this executable.
             # -s is important! Can otherwise leak packages to other environments.
-            return os.path.join("linux-64", id_, "bin/python -s")
+            return os.path.join("$MF_ARCH", id_, "bin/python -s")
         else:
             # for @conda/@pypi(disabled=True).
             return super().executable(step_name, default)
@@ -315,7 +314,6 @@ class CondaEnvironment(MetaflowEnvironment):
         # 5. All resolved packages (Conda or PyPI) are cached
         # 6. PyPI packages are only installed for local platform
 
-        # Resolve `linux-64` Conda environments if @batch or @kubernetes are in play
         target_platform = conda_platform()
         for decorator in step.decorators:
             # NOTE: Keep the list of supported decorator names for backward compatibility purposes.
@@ -329,7 +327,6 @@ class CondaEnvironment(MetaflowEnvironment):
                 "snowpark",
                 "slurm",
             ]:
-                # TODO: Support arm architectures
                 target_platform = getattr(decorator, "target_platform", "linux-64")
                 break
 
@@ -424,15 +421,18 @@ class CondaEnvironment(MetaflowEnvironment):
         if id_:
             return [
                 "echo 'Bootstrapping virtual environment...'",
+                "flush_mflogs",
                 # We have to prevent the tracing module from loading,
                 # as the bootstrapping process uses the internal S3 client which would fail to import tracing
                 # due to the required dependencies being bundled into the conda environment,
                 # which is yet to be initialized at this point.
-                'DISABLE_TRACING=True python -m metaflow.plugins.pypi.bootstrap "%s" %s "%s" linux-64'
+                'DISABLE_TRACING=True python -m metaflow.plugins.pypi.bootstrap "%s" %s "%s"'
                 % (self.flow.name, id_, self.datastore_type),
                 "echo 'Environment bootstrapped.'",
+                "flush_mflogs",
                 # To avoid having to install micromamba in the PATH in micromamba.py, we add it to the PATH here.
                 "export PATH=$PATH:$(pwd)/micromamba/bin",
+                "export MF_ARCH=$(case $(uname)/$(uname -m) in Darwin/arm64)echo osx-arm64;;Darwin/*)echo osx-64;;Linux/aarch64)echo linux-aarch64;;*)echo linux-64;;esac)",
             ]
         else:
             # for @conda/@pypi(disabled=True).

metaflow/plugins/pypi/parsers.py

@@ -0,0 +1,268 @@
+# this file can be overridden by extensions as is (e.g. metaflow-nflx-extensions)
+from metaflow.exception import MetaflowException
+
+
+class ParserValueError(MetaflowException):
+    headline = "Value error"
+
+
+def requirements_txt_parser(content: str):
+    """
+    Parse non-comment lines from a requirements.txt file as strictly valid
+    PEP 508 requirements.
+
+    Recognizes direct references (e.g. "my_lib @ git+https://..."), extras
+    (e.g. "requests[security]"), and version specifiers (e.g. "==2.0"). If
+    the package name is "python", its specifier is stored in the "python"
+    key instead of "packages".
+
+    Parameters
+    ----------
+    content : str
+        Contents of a requirements.txt file.
+
+    Returns
+    -------
+    dict
+        A dictionary with two keys:
+            - "packages": dict(str -> str)
+              Mapping from package name (plus optional extras/references) to a
+              version specifier string.
+            - "python": str or None
+              The Python version constraints if present, otherwise None.
+
+    Raises
+    ------
+     ParserValueError
+        If a requirement line is invalid PEP 508 or if environment markers are
+        detected, or if multiple Python constraints are specified.
+    """
+    import re
+    from metaflow._vendor.packaging.requirements import Requirement, InvalidRequirement
+
+    parsed = {"packages": {}, "python": None}
+
+    inline_comment_pattern = re.compile(r"\s+#.*$")
+    for line in content.splitlines():
+        line = line.strip()
+
+        # support Rye lockfiles by skipping lines not compliant with requirements
+        if line == "-e file:.":
+            continue
+
+        if not line or line.startswith("#"):
+            continue
+
+        line = inline_comment_pattern.sub("", line).strip()
+        if not line:
+            continue
+
+        try:
+            req = Requirement(line)
+        except InvalidRequirement:
+            raise ParserValueError(f"Not a valid PEP 508 requirement: '{line}'")
+
+        if req.marker is not None:
+            raise ParserValueError(
+                "Environment markers (e.g. 'platform_system==\"Linux\"') "
+                f"are not supported for line: '{line}'"
+            )
+
+        dep_key = req.name
+        if req.extras:
+            dep_key += f"[{','.join(req.extras)}]"
+        if req.url:
+            dep_key += f"@{req.url}"
+
+        dep_spec = str(req.specifier).lstrip(" =")
+
+        if req.name.lower() == "python":
+            if parsed["python"] is not None and dep_spec:
+                raise ParserValueError(
+                    f"Multiple Python version specs not allowed: '{line}'"
+                )
+            parsed["python"] = dep_spec or None
+        else:
+            parsed["packages"][dep_key] = dep_spec
+
+    return parsed
+
+
+def pyproject_toml_parser(content: str):
+    """
+    Parse a pyproject.toml file per PEP 621.
+
+    Reads the 'requires-python' and 'dependencies' fields from the "[project]" section.
+    Each dependency line must be a valid PEP 508 requirement. If the package name is
+    "python", its specifier is stored in the "python" key instead of "packages".
+
+    Parameters
+    ----------
+    content : str
+        Contents of a pyproject.toml file.
+
+    Returns
+    -------
+    dict
+        A dictionary with two keys:
+            - "packages": dict(str -> str)
+              Mapping from package name (plus optional extras/references) to a
+              version specifier string.
+            - "python": str or None
+              The Python version constraints if present, otherwise None.
+
+    Raises
+    ------
+    RuntimeError
+        If no TOML library (tomllib in Python 3.11+ or tomli in earlier versions) is found.
+     ParserValueError
+        If a dependency is not valid PEP 508, if environment markers are used, or if
+        multiple Python constraints are specified.
+    """
+    try:
+        import tomllib as toml  # Python 3.11+
+    except ImportError:
+        try:
+            import tomli as toml  # Python < 3.11 (requires "tomli" package)
+        except ImportError:
+            raise RuntimeError(
+                "Could not import a TOML library. For Python <3.11, please install 'tomli'."
+            )
+    from metaflow._vendor.packaging.requirements import Requirement, InvalidRequirement
+
+    data = toml.loads(content)
+
+    project = data.get("project", {})
+    requirements = project.get("dependencies", [])
+    requires_python = project.get("requires-python")
+
+    parsed = {"packages": {}, "python": None}
+
+    if requires_python is not None:
+        # If present, store verbatim; note that PEP 621 does not necessarily
+        # require "python" to be a dependency in the usual sense.
+        # Example: "requires-python" = ">=3.7,<4"
+        parsed["python"] = requires_python.lstrip("=").strip()
+
+    for dep_line in requirements:
+        dep_line_stripped = dep_line.strip()
+        try:
+            req = Requirement(dep_line_stripped)
+        except InvalidRequirement:
+            raise ParserValueError(
+                f"Not a valid PEP 508 requirement: '{dep_line_stripped}'"
+            )
+
+        if req.marker is not None:
+            raise ParserValueError(
+                f"Environment markers not supported for line: '{dep_line_stripped}'"
+            )
+
+        dep_key = req.name
+        if req.extras:
+            dep_key += f"[{','.join(req.extras)}]"
+        if req.url:
+            dep_key += f"@{req.url}"
+
+        dep_spec = str(req.specifier).lstrip("=")
+
+        if req.name.lower() == "python":
+            if parsed["python"] is not None and dep_spec:
+                raise ParserValueError(
+                    f"Multiple Python version specs not allowed: '{dep_line_stripped}'"
+                )
+            parsed["python"] = dep_spec or None
+        else:
+            parsed["packages"][dep_key] = dep_spec
+
+    return parsed
+
+
+def conda_environment_yml_parser(content: str):
+    """
+    Parse a minimal environment.yml file under strict assumptions.
+
+    The file must contain a 'dependencies:' line, after which each dependency line
+    appears with a '- ' prefix. Python can appear as 'python=3.9', etc.; other
+    packages as 'numpy=1.21.2' or simply 'numpy'. Non-compliant lines raise  ParserValueError.
+
+    Parameters
+    ----------
+    content : str
+        Contents of a environment.yml file.
+
+    Returns
+    -------
+    dict
+        A dictionary with keys:
+        {
+            "packages": dict(str -> str),
+            "python": str or None
+        }
+
+    Raises
+    ------
+     ParserValueError
+        If the file has malformed lines or unsupported sections.
+    """
+    import re
+
+    packages = {}
+    python_version = None
+
+    inside_dependencies = False
+
+    # Basic pattern for lines like "numpy=1.21.2"
+    # Group 1: package name
+    # Group 2: optional operator + version (could be "=1.21.2", "==1.21.2", etc.)
+    line_regex = re.compile(r"^([A-Za-z0-9_\-\.]+)(\s*[=<>!~].+\s*)?$")
+    inline_comment_pattern = re.compile(r"\s+#.*$")
+
+    for line in content.splitlines():
+        line = line.strip()
+        if not line or line.startswith("#"):
+            continue
+
+        line = inline_comment_pattern.sub("", line).strip()
+        if not line:
+            continue
+
+        if line.lower().startswith("dependencies:"):
+            inside_dependencies = True
+            continue
+
+        if inside_dependencies and not line.startswith("-"):
+            inside_dependencies = False
+            continue
+
+        if not inside_dependencies:
+            continue
+
+        dep_line = line.lstrip("-").strip()
+        if dep_line.endswith(":"):
+            raise ParserValueError(
+                f"Unsupported subsection '{dep_line}' in environment.yml."
+            )
+
+        match = line_regex.match(dep_line)
+        if not match:
+            raise ParserValueError(
+                f"Line '{dep_line}' is not a valid conda package specifier."
+            )
+
+        pkg_name, pkg_version_part = match.groups()
+        version_spec = pkg_version_part.strip() if pkg_version_part else ""
+
+        if version_spec.startswith("="):
+            version_spec = version_spec.lstrip("=").strip()
+
+        if pkg_name.lower() == "python":
+            if python_version is not None and version_spec:
+                raise ParserValueError(
+                    f"Multiple Python version specs detected: '{dep_line}'"
+                )
+            python_version = version_spec
+        else:
+            packages[pkg_name] = version_spec
+
+    return {"packages": packages, "python": python_version}

metaflow/plugins/pypi/utils.py

@@ -72,6 +72,24 @@ def pip_tags(python_version, mamba_platform):
             )
         ]
         platforms.append("linux_x86_64")
+    elif mamba_platform == "linux-aarch64":
+        platforms = [
+            "manylinux%s_aarch64" % s
+            for s in (
+                "2014",
+                "_2_17",
+                "_2_18",
+                "_2_19",
+                "_2_20",
+                "_2_21",
+                "_2_23",
+                "_2_24",
+                "_2_25",
+                "_2_26",
+                "_2_27",
+            )
+        ]
+        platforms.append("linux_aarch64")
     elif mamba_platform == "osx-64":
         platforms = tags.mac_platforms(arch="x86_64")
     elif mamba_platform == "osx-arm64":

metaflow/version.py

@@ -1 +1 @@
-metaflow_version = "2.15.0"
+metaflow_version = "2.15.1"

{metaflow-2.15.0.data → metaflow-2.15.1.data}/data/share/metaflow/devtools/Makefile

@@ -250,8 +250,7 @@ shell: setup-tilt
 				"$$user_shell" -i; \
 		fi'
 
-# TODO: This can be done away with in a while since we now have metaflow-dev command.
-#
+
 # @echo '$(MAKE_CMD) create-dev-shell' >> $(DEVTOOLS_DIR)/start.sh
 # @echo 'rm -f /tmp/metaflow-devshell-*' >> $(DEVTOOLS_DIR)/start.sh
 create-dev-shell: setup-tilt

{metaflow-2.15.0.data → metaflow-2.15.1.data}/data/share/metaflow/devtools/Tiltfile

@@ -99,6 +99,7 @@ if "minio" in enabled_components:
         set=[
             'rootUser=rootuser',
             'rootPassword=rootpass123',
+            # TODO: perturb the bucket name to avoid conflicts
             'buckets[0].name=metaflow-test',
             'buckets[0].policy=none',
             'buckets[0].purge=false',
@@ -600,6 +601,18 @@ if "ui" in enabled_components:
     config_resources.append('metaflow-ui')
     config_resources.append('metaflow-ui-static')
 
+cmd = '''
+ARCH=$(kubectl get nodes -o jsonpath='{.items[0].status.nodeInfo.architecture}')
+case "$ARCH" in
+  arm64)   echo linux-aarch64 ;;
+  amd64)   echo linux-64 ;;
+  *)       echo linux-64 ;;
+esac
+'''
+
+# For @conda/@pypi emulation
+metaflow_config["METAFLOW_KUBERNETES_CONDA_ARCH"] = str(local(cmd)).strip()
+
 local_resource(
     name="generate-configs",
     cmd=write_config_files(),

{metaflow-2.15.0.dist-info → metaflow-2.15.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: metaflow
-Version: 2.15.0
+Version: 2.15.1
 Summary: Metaflow: More AI and ML, Less Engineering
 Author: Metaflow Developers
 Author-email: help@metaflow.org
@@ -26,7 +26,7 @@ License-File: LICENSE
 Requires-Dist: requests
 Requires-Dist: boto3
 Provides-Extra: stubs
-Requires-Dist: metaflow-stubs==2.15.0; extra == "stubs"
+Requires-Dist: metaflow-stubs==2.15.1; extra == "stubs"
 Dynamic: author
 Dynamic: author-email
 Dynamic: classifier

{metaflow-2.15.0.dist-info → metaflow-2.15.1.dist-info}/RECORD

@@ -1,5 +1,5 @@
 metaflow/R.py,sha256=CqVfIatvmjciuICNnoyyNGrwE7Va9iXfLdFbQa52hwA,3958
-metaflow/__init__.py,sha256=fbhdWiWnEoAX4KnzRHMY_iQcT-uYlMWhzrXPKvK0i5g,5832
+metaflow/__init__.py,sha256=6ws8LpYz5HwIrc1J4BK-2Hxs0okGLFQCm31bHFf2H8E,5978
 metaflow/cards.py,sha256=IbRmredvmFEU0V6JL7DR8wCESwVmmZJubr6x24bo7U4,442
 metaflow/cli.py,sha256=RU-yXpT-Lfl3xGyFNtL742e9KEqcRxEnQ-4mwXrXhvo,20928
 metaflow/cli_args.py,sha256=hDsdWdRmfXYifVGq6b6FDfgoWxtIG2nr_lU6EBV0Pnk,3584
@@ -16,10 +16,10 @@ metaflow/includefile.py,sha256=kWKDSlzVcRVNGG9PV5eB3o2ynrzqhVsfaLtkqjshn7Q,20948
 metaflow/info_file.py,sha256=wtf2_F0M6dgiUu74AFImM8lfy5RrUw5Yj7Rgs2swKRY,686
 metaflow/integrations.py,sha256=LlsaoePRg03DjENnmLxZDYto3NwWc9z_PtU6nJxLldg,1480
 metaflow/lint.py,sha256=x4p6tnRzYqNNniCGXyrUW0WuYfTUgnaOMRivxvnxask,11661
-metaflow/metaflow_config.py,sha256=jCyJhKrBePqygpfZ4C-gmU14FlTesvFmmimnTP3t_U8,23433
+metaflow/metaflow_config.py,sha256=oLbF4ZOfdejRBbemL_9NmFo2G2iAdTuUgbd7vNxV2lg,23567
 metaflow/metaflow_config_funcs.py,sha256=5GlvoafV6SxykwfL8D12WXSfwjBN_NsyuKE_Q3gjGVE,6738
 metaflow/metaflow_current.py,sha256=pfkXmkyHeMJhxIs6HBJNBEaBDpcl5kz9Wx5mW6F_3qo,7164
-metaflow/metaflow_environment.py,sha256=gs_UpYpuOKEEfFM0z0tnwje7zAVsQ5Ck7Dp2M9_1utQ,8065
+metaflow/metaflow_environment.py,sha256=e5BOkA7VdpjseI4HUkm_pR74NVJRNADL20LIQL4W1vU,8139
 metaflow/metaflow_profile.py,sha256=jKPEW-hmAQO-htSxb9hXaeloLacAh41A35rMZH6G8pA,418
 metaflow/metaflow_version.py,sha256=duhIzfKZtcxMVMs2uiBqBvUarSHJqyWDwMhaBOQd_g0,7491
 metaflow/monitor.py,sha256=T0NMaBPvXynlJAO_avKtk8OIIRMyEuMAyF8bIp79aZU,5323
@@ -36,7 +36,7 @@ metaflow/tuple_util.py,sha256=_G5YIEhuugwJ_f6rrZoelMFak3DqAR2tt_5CapS1XTY,830
 metaflow/unbounded_foreach.py,sha256=p184WMbrMJ3xKYHwewj27ZhRUsSj_kw1jlye5gA9xJk,387
 metaflow/util.py,sha256=mJBkV5tShIyCsLDeM1zygQGeciQVMrVPm_qI8Oi33G0,14656
 metaflow/vendor.py,sha256=FchtA9tH22JM-eEtJ2c9FpUdMn8sSb1VHuQS56EcdZk,5139
-metaflow/version.py,sha256=IWpJOWnkdfrf92MEtc6EPVbrvSldtZ2YIiIqxV_gwJI,28
+metaflow/version.py,sha256=VW8RqjdxyB49j8TFdvknh5vAqtyjJ9j4ApSZP3QA-zs,28
 metaflow/_vendor/__init__.py,sha256=y_CiwUD3l4eAKvTVDZeqgVujMy31cAM1qjAB-HfI-9s,353
 metaflow/_vendor/typing_extensions.py,sha256=0nUs5p1A_UrZigrAVBoOEM6TxU37zzPDUtiij1ZwpNc,110417
 metaflow/_vendor/zipp.py,sha256=ajztOH-9I7KA_4wqDYygtHa6xUBVZgFpmZ8FE74HHHI,8425
@@ -121,12 +121,13 @@ metaflow/client/core.py,sha256=Cca6HbK-UBO72aELfFJxsl85ylYHHlCAd-uJP-lEepQ,83689
 metaflow/client/filecache.py,sha256=Wy0yhhCqC1JZgebqi7z52GCwXYnkAqMZHTtxThvwBgM,15229
 metaflow/cmd/__init__.py,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
 metaflow/cmd/configure_cmd.py,sha256=o-DKnUf2FBo_HiMVyoyzQaGBSMtpbEPEdFTQZ0hkU-k,33396
-metaflow/cmd/main_cli.py,sha256=H0UC-jiZjThHZgQeMyNZh3raSDDyUTvMspYUqKFnNSU,2955
+metaflow/cmd/main_cli.py,sha256=LSehmMjkWojAN1XTtqW6S51ZpGNAdW4_VK5S7qH8-Ts,2982
 metaflow/cmd/make_wrapper.py,sha256=NFpSdESs4Ks9xeurmYB5VUyYplhNcONDZJcUP2cf8-8,494
 metaflow/cmd/tutorials_cmd.py,sha256=8FdlKkicTOhCIDKcBR5b0Oz6giDvS-EMY3o9skIrRqw,5156
 metaflow/cmd/util.py,sha256=jS_0rUjOnGGzPT65fzRLdGjrYAOOLA4jU2S0HJLV0oc,406
+metaflow/cmd/code/__init__.py,sha256=VO4dNM9M9LHYy5nTgEiJvCV1RBl8lpDlYGJm6GIcaBA,7413
 metaflow/cmd/develop/__init__.py,sha256=p1Sy8yU1MEKSrH5ttOWOZvNcI1qYu6J6jghdTHwPgOw,689
-metaflow/cmd/develop/stub_generator.py,sha256=vhhEAPkHN-9wZYW4rqnMjVsjwqm4Avn_inVx-wC6DKE,65164
+metaflow/cmd/develop/stub_generator.py,sha256=bo2yWe0kvCZ-3arEFe9eAnPN-h8oNNPcQsDwsL350UM,65217
 metaflow/cmd/develop/stubs.py,sha256=JX2qNZDvG0upvPueAcLhoR_zyLtRranZMwY05tLdpRQ,11884
 metaflow/datastore/__init__.py,sha256=VxP6ddJt3rwiCkpiSfAhyVkUCOe1pgZZsytVEJzFmSQ,155
 metaflow/datastore/content_addressed_store.py,sha256=6T7tNqL29kpmecyMLHF35RhoSBOb-OZcExnsB65AvnI,7641
@@ -145,7 +146,7 @@ metaflow/metadata_provider/__init__.py,sha256=FZNSnz26VB_m18DQG8mup6-Gfl7r1U6lRM
 metaflow/metadata_provider/heartbeat.py,sha256=42mQo6wOHdFuaCh426uV6Kn8swe7e5I3gqA_G7cI_LA,3127
 metaflow/metadata_provider/metadata.py,sha256=meO4Fhxu7tbMUGwasYb9_AtL06fwrrXKKjIK7KRWZDs,27093
 metaflow/metadata_provider/util.py,sha256=lYoQKbqoTM1iZChgyVWN-gX-HyM9tt9bXEMJexY9XmM,1723
-metaflow/mflog/__init__.py,sha256=9iMMn2xYB0oaDXXcInxa9AdDqeVBeiJeB3klnqGkyL0,5983
+metaflow/mflog/__init__.py,sha256=TkR9ny_JYvNCWJTdLiHsbLSLc9cUvzAzpDuHLdG8nkA,6020
 metaflow/mflog/mflog.py,sha256=VebXxqitOtNAs7VJixnNfziO_i_urG7bsJ5JiB5IXgY,4370
 metaflow/mflog/save_logs.py,sha256=4p1OwozsHJBslOzAf0wUq2XPMNpEOZWM68MgWzh_jJY,2330
 metaflow/mflog/save_logs_periodically.py,sha256=2Uvk9hi-zlCqXxOQoXmmjH1SCugfw6eG6w70WgfI-ho,1256
@@ -292,23 +293,24 @@ metaflow/plugins/kubernetes/kube_utils.py,sha256=jdFMGbEmIow-oli26v31W9CmbZXigx0
 metaflow/plugins/kubernetes/kubernetes.py,sha256=g_E5jBhoMSDzGFnP5PDQiVPPllUr1wulVcG3tz247G8,29668
 metaflow/plugins/kubernetes/kubernetes_cli.py,sha256=PY-aMaVelHddHq3jqEEu9cvNl7xEjT2lFFADN9dXMkw,13918
 metaflow/plugins/kubernetes/kubernetes_client.py,sha256=tuvXP-QKpdeSmzVolB2R_TaacOr5DIb0j642eKcjsiM,6491
-metaflow/plugins/kubernetes/kubernetes_decorator.py,sha256=Yg32f1VWPPvV2pECufBG4VqyZw_QNY39R836G8UVLKw,30844
+metaflow/plugins/kubernetes/kubernetes_decorator.py,sha256=ydJ6WyatCSwUzq_xGcjG1nCp1yRNhSyh09z0Y1cwDNI,30896
 metaflow/plugins/kubernetes/kubernetes_job.py,sha256=pO9ExyAVCDoAoWFn9oFcos2aa0MQk4_D61O-T4E10E8,31826
-metaflow/plugins/kubernetes/kubernetes_jobsets.py,sha256=9kU43eE5IvIa7y-POzBdxnJOazWsedKhwQ51Tu1HN_A,42471
+metaflow/plugins/kubernetes/kubernetes_jobsets.py,sha256=XjduAYY_H-jX3lGK6b-jYt6QnDvC2ac3qIAgr-iDsCQ,42543
 metaflow/plugins/kubernetes/spot_metadata_cli.py,sha256=an0nWCxgflmqIPBCBrlb4m3DereDFFJBLt-KKhqcHc8,1670
 metaflow/plugins/kubernetes/spot_monitor_sidecar.py,sha256=zrWU-smQwPnL6MBHmzTxWyEA00R6iKKQbhhy50xFwQ8,3832
 metaflow/plugins/metadata_providers/__init__.py,sha256=AbpHGcgLb-kRsJGnwFEktk7uzpZOCcBY74-YBdrKVGs,1
 metaflow/plugins/metadata_providers/local.py,sha256=Z0CXaGZJbAkj4II3WspJi-uCCtShH64yaXZQ5i9Ym7g,24390
 metaflow/plugins/metadata_providers/service.py,sha256=K0Ym6lcmegX6wBC5uZbeAFQJSDFc8e6DzJiCB1VIqjc,22554
 metaflow/plugins/pypi/__init__.py,sha256=0YFZpXvX7HCkyBFglatual7XGifdA1RwC3U4kcizyak,1037
-metaflow/plugins/pypi/bootstrap.py,sha256=XAz832qSLFxIXW6SP02N8PQ_7CKiqrCfirkE80Iwarc,14881
+metaflow/plugins/pypi/bootstrap.py,sha256=SNONquX6QnTbu7htmhaQeVeZ2ofaFaUCDScRIrTTERc,14718
 metaflow/plugins/pypi/conda_decorator.py,sha256=piFcE4uGmWhhbGlxMK0GHd7BGEyqy6r9BFy8Mjoi80Q,15937
-metaflow/plugins/pypi/conda_environment.py,sha256=d5BAiY_aJJdlJ5h3N5nGSDmVoOY-8BVKqEbA5nrCpCY,22113
+metaflow/plugins/pypi/conda_environment.py,sha256=yeTPGuu38EQ8aYzXUbc6cLc9b2NMLLEnxS-C73PcDHk,22250
 metaflow/plugins/pypi/micromamba.py,sha256=LLJ2dGGOEyld07W8iI6dtE01h2Y1PQnBhU-dMBssZ3c,16502
+metaflow/plugins/pypi/parsers.py,sha256=gpOOG2Ph95wI73MWCAi7XjpK0gYhv5k5YIGBs73QPuE,8556
 metaflow/plugins/pypi/pip.py,sha256=H0cIy8odpZ-JTn4SwF0b74tuC3uRU7X8TdAQJ2kODG8,13971
 metaflow/plugins/pypi/pypi_decorator.py,sha256=ybNgo-T5Z_0W2KNuED0pdjyI0qygZ4a1MXAzKqdHt_E,7250
 metaflow/plugins/pypi/pypi_environment.py,sha256=FYMg8kF3lXqcLfRYWD83a9zpVjcoo_TARqMGZ763rRk,230
-metaflow/plugins/pypi/utils.py,sha256=855aSATi-qPhFs5OFV5dl03RDNDr5tUrPebloygrWnU,2984
+metaflow/plugins/pypi/utils.py,sha256=glfXN0Do8HleB8YE9LE9Pb1tIBivqdPGmS0sUtO1e7k,3451
 metaflow/plugins/secrets/__init__.py,sha256=mhJaN2eMS_ZZVewAMR2E-JdP5i0t3v9e6Dcwd-WpruE,310
 metaflow/plugins/secrets/inline_secrets_provider.py,sha256=EChmoBGA1i7qM3jtYwPpLZDBybXLergiDlN63E0u3x8,294
 metaflow/plugins/secrets/secrets_decorator.py,sha256=s-sFzPWOjahhpr5fMj-ZEaHkDYAPTO0isYXGvaUwlG8,11273
@@ -361,12 +363,12 @@ metaflow/user_configs/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3h
 metaflow/user_configs/config_decorators.py,sha256=qCKVAvd0NKgaCxQ2OThes5-DYHXq6A1HqURubYNeFdw,20481
 metaflow/user_configs/config_options.py,sha256=m6jccSpzI4qUJ7vyYkYBIf8G3V0Caunxg_k7zg4Zlqg,21067
 metaflow/user_configs/config_parameters.py,sha256=oeJGVKu1ao_YQX6Lg6P2FEv5k5-_F4sARLlVpTW9ezM,15502
-metaflow-2.15.0.data/data/share/metaflow/devtools/Makefile,sha256=x9Q2FsScc9XQa0uVV2oNpA3VHwet_6oc8aQN0ztbM2Q,12907
-metaflow-2.15.0.data/data/share/metaflow/devtools/Tiltfile,sha256=ednswaJXxyH4wRWPNQZMzb5Kg1TiukHUNXgUh_DP8mU,20016
-metaflow-2.15.0.data/data/share/metaflow/devtools/pick_services.sh,sha256=DCnrMXwtApfx3B4S-YiZESMyAFHbXa3VuNL0MxPLyiE,2196
-metaflow-2.15.0.dist-info/LICENSE,sha256=nl_Lt5v9VvJ-5lWJDT4ddKAG-VZ-2IaLmbzpgYDz2hU,11343
-metaflow-2.15.0.dist-info/METADATA,sha256=cvzxosgu6SxOBgmkE_UHfdyJEIkvadf54EsYe3npi8k,6118
-metaflow-2.15.0.dist-info/WHEEL,sha256=3HfeesdN7jshHPzN8HJ8UeFRlEd6ApplqndzbPTvPl8,109
-metaflow-2.15.0.dist-info/entry_points.txt,sha256=RvEq8VFlgGe_FfqGOZi0D7ze1hLD0pAtXeNyGfzc_Yc,103
-metaflow-2.15.0.dist-info/top_level.txt,sha256=v1pDHoWaSaKeuc5fKTRSfsXCKSdW1zvNVmvA-i0if3o,9
-metaflow-2.15.0.dist-info/RECORD,,
+metaflow-2.15.1.data/data/share/metaflow/devtools/Makefile,sha256=6B1iQvyfhhukbDnLnzZmadYi9tmyr538hVmKiBIy_Ew,12820
+metaflow-2.15.1.data/data/share/metaflow/devtools/Tiltfile,sha256=fpPW_0sDxHL60-9nld5OgV0YO30IPl9luAjukTsl-9o,20391
+metaflow-2.15.1.data/data/share/metaflow/devtools/pick_services.sh,sha256=DCnrMXwtApfx3B4S-YiZESMyAFHbXa3VuNL0MxPLyiE,2196
+metaflow-2.15.1.dist-info/LICENSE,sha256=nl_Lt5v9VvJ-5lWJDT4ddKAG-VZ-2IaLmbzpgYDz2hU,11343
+metaflow-2.15.1.dist-info/METADATA,sha256=uZRHK4OO4RES9LB_ZJGiL0oFxq3Z_VYZZWqnEeMJhnw,6118
+metaflow-2.15.1.dist-info/WHEEL,sha256=rF4EZyR2XVS6irmOHQIJx2SUqXLZKRMUrjsg8UwN-XQ,109
+metaflow-2.15.1.dist-info/entry_points.txt,sha256=RvEq8VFlgGe_FfqGOZi0D7ze1hLD0pAtXeNyGfzc_Yc,103
+metaflow-2.15.1.dist-info/top_level.txt,sha256=v1pDHoWaSaKeuc5fKTRSfsXCKSdW1zvNVmvA-i0if3o,9
+metaflow-2.15.1.dist-info/RECORD,,

{metaflow-2.15.0.dist-info → metaflow-2.15.1.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (75.8.1)
+Generator: setuptools (75.8.2)
 Root-Is-Purelib: true
 Tag: py2-none-any
 Tag: py3-none-any