ob-metaflow 2.12.36.3__py2.py3-none-any.whl → 2.13.0.1__py2.py3-none-any.whl

metaflow/plugins/aws/secrets_manager/aws_secrets_manager_secrets_provider.py

@@ -50,24 +50,27 @@ class AwsSecretsManagerSecretsProvider(SecretsProvider):
         The secret payload from AWS is EITHER a string OR a binary blob.
 
         If the secret contains a string payload ("SecretString"):
-        - if the `parse_secret_string_as_json` option is True (default):
+        - if the `json` option is True (default):
             {SecretString} will be parsed as a JSON. If successfully parsed, AND the JSON contains a
             top-level object, each entry K/V in the object will also be converted to an entry in the result. V will
             always be casted to a string (if not already a string).
-        - If `parse_secret_string_as_json` option is False:
-            {SecretString} will be returned as a single entry in the result, with the key being the secret_id.
+        - If `json` option is False:
+            {SecretString} will be returned as a single entry in the result, where the key is either:
+                - the `secret_id`, OR
+                - the value set by `options={"env_var_name": custom_env_var_name}`.
 
-        Otherwise, the secret contains a binary blob payload ("SecretBinary"). In this case
-        - The result dic contains '{SecretName}': '{SecretBinary}', where {SecretBinary} is a base64-encoded string
+        Otherwise, if the secret contains a binary blob payload ("SecretBinary"):
+        - The result dict contains '{SecretName}': '{SecretBinary}', where {SecretBinary} is a base64-encoded string.
 
-        All keys in the result are sanitized to be more valid environment variable names. This is done on a best effort
+        All keys in the result are sanitized to be more valid environment variable names. This is done on a best-effort
         basis. Further validation is expected to be done by the invoking @secrets decorator itself.
 
-        :param secret_id: ARN or friendly name of the secret
-        :param options: unused
-        :param role: AWS IAM Role ARN to assume before reading the secret
-        :return: dict of environment variables. All keys and values are strings.
+        :param secret_id: ARN or friendly name of the secret.
+        :param options: Dictionary of additional options. E.g., `options={"env_var_name": custom_env_var_name}`.
+        :param role: AWS IAM Role ARN to assume before reading the secret.
+        :return: Dictionary of environment variables. All keys and values are strings.
         """
+
         import botocore
         from metaflow.plugins.aws.aws_client import get_aws_client
 

metaflow/plugins/aws/step_functions/step_functions.py

@@ -18,6 +18,7 @@ from metaflow.metaflow_config import (
     SFN_S3_DISTRIBUTED_MAP_OUTPUT_PATH,
 )
 from metaflow.parameters import deploy_time_eval
+from metaflow.user_configs.config_options import ConfigInput
 from metaflow.util import dict_to_cli_options, to_pascalcase
 
 from ..batch.batch import Batch
@@ -71,6 +72,7 @@ class StepFunctions(object):
         self.username = username
         self.max_workers = max_workers
         self.workflow_timeout = workflow_timeout
+        self.config_parameters = self._process_config_parameters()
 
         # https://aws.amazon.com/blogs/aws/step-functions-distributed-map-a-serverless-solution-for-large-scale-parallel-data-processing/
         self.use_distributed_map = use_distributed_map
@@ -485,6 +487,10 @@ class StepFunctions(object):
                     "case-insensitive." % param.name
                 )
             seen.add(norm)
+            # NOTE: We skip config parameters as these do not have dynamic values,
+            # and need to be treated differently.
+            if param.IS_CONFIG_PARAMETER:
+                continue
 
             is_required = param.kwargs.get("required", False)
             # Throw an exception if a schedule is set for a flow with required
@@ -501,6 +507,27 @@ class StepFunctions(object):
             parameters.append(dict(name=param.name, value=value))
         return parameters
 
+    def _process_config_parameters(self):
+        parameters = []
+        seen = set()
+        for var, param in self.flow._get_parameters():
+            if not param.IS_CONFIG_PARAMETER:
+                continue
+            # Throw an exception if the parameter is specified twice.
+            norm = param.name.lower()
+            if norm in seen:
+                raise MetaflowException(
+                    "Parameter *%s* is specified twice. "
+                    "Note that parameter names are "
+                    "case-insensitive." % param.name
+                )
+            seen.add(norm)
+
+            parameters.append(
+                dict(name=param.name, kv_name=ConfigInput.make_key_name(param.name))
+            )
+        return parameters
+
     def _batch(self, node):
         attrs = {
             # metaflow.user is only used for setting the AWS Job Name.
@@ -747,6 +774,11 @@ class StepFunctions(object):
         metaflow_version["production_token"] = self.production_token
         env["METAFLOW_VERSION"] = json.dumps(metaflow_version)
 
+        # map config values
+        cfg_env = {param["name"]: param["kv_name"] for param in self.config_parameters}
+        if cfg_env:
+            env["METAFLOW_FLOW_CONFIG_VALUE"] = json.dumps(cfg_env)
+
         # Set AWS DynamoDb Table Name for state tracking for for-eaches.
         # There are three instances when metaflow runtime directly interacts
         # with AWS DynamoDB.

metaflow/plugins/aws/step_functions/step_functions_cli.py

@@ -326,6 +326,7 @@ def make_flow(
 
     # Attach AWS Batch decorator to the flow
     decorators._attach_decorators(obj.flow, [BatchDecorator.name])
+    decorators._init(obj.flow)
     decorators._init_step_decorators(
         obj.flow, obj.graph, obj.environment, obj.flow_datastore, obj.logger
     )

metaflow/plugins/aws/step_functions/step_functions_deployer_objects.py

@@ -46,6 +46,7 @@ class StepFunctionsTriggeredRun(TriggeredRun):
         )
 
         command_obj = self.deployer.spm.get(pid)
+        command_obj.sync_wait()
         return command_obj.process.returncode == 0
 
 
@@ -174,6 +175,7 @@ class StepFunctionsDeployedFlow(DeployedFlow):
         )
 
         command_obj = self.deployer.spm.get(pid)
+        command_obj.sync_wait()
         return command_obj.process.returncode == 0
 
     def trigger(self, **kwargs) -> StepFunctionsTriggeredRun:
@@ -217,6 +219,7 @@ class StepFunctionsDeployedFlow(DeployedFlow):
                 attribute_file_fd, command_obj, self.deployer.file_read_timeout
             )
 
+            command_obj.sync_wait()
             if command_obj.process.returncode == 0:
                 return StepFunctionsTriggeredRun(
                     deployer=self.deployer, content=content

metaflow/plugins/cards/card_creator.py

@@ -122,6 +122,7 @@ class CardCreator:
             executable,
             sys.argv[0],
         ]
+
         cmd += self._top_level_options + [
             "card",
             "create",

metaflow/plugins/cards/card_decorator.py

@@ -1,13 +1,16 @@
+import json
+import os
+import re
+import tempfile
+
 from metaflow.decorators import StepDecorator
 from metaflow.metaflow_current import current
+from metaflow.user_configs.config_options import ConfigInput
+from metaflow.user_configs.config_parameters import dump_config_values
 from metaflow.util import to_unicode
+
 from .component_serializer import CardComponentCollector, get_card_class
 from .card_creator import CardCreator
-
-
-# from metaflow import get_metadata
-import re
-
 from .exception import CARD_ID_PATTERN, TYPE_CHECK_REGEX
 
 ASYNC_TIMEOUT = 30
@@ -111,6 +114,14 @@ class CardDecorator(StepDecorator):
         self._logger = logger
         self.card_options = None
 
+        # We check for configuration options. We do this here before they are
+        # converted to properties.
+        self._config_values = [
+            (config.name, ConfigInput.make_key_name(config.name))
+            for _, config in flow._get_parameters()
+            if config.IS_CONFIG_PARAMETER
+        ]
+
         self.card_options = self.attributes["options"]
 
         evt_name = "step-init"
@@ -146,6 +157,18 @@ class CardDecorator(StepDecorator):
         self._task_datastore = task_datastore
         self._metadata = metadata
 
+        # If we have configs, we need to dump them to a file so we can re-use them
+        # when calling the card creation subprocess.
+        if self._config_values:
+            with tempfile.NamedTemporaryFile(
+                mode="w", encoding="utf-8", delete=False
+            ) as config_file:
+                config_value = dump_config_values(flow)
+                json.dump(config_value, config_file)
+                self._config_file_name = config_file.name
+        else:
+            self._config_file_name = None
+
         card_type = self.attributes["type"]
         card_class = get_card_class(card_type)
 
@@ -179,7 +202,7 @@ class CardDecorator(StepDecorator):
         # we need to ensure that `current.card` has `CardComponentCollector` instantiated only once.
         if not self._is_event_registered("pre-step"):
             self._register_event("pre-step")
-            self._set_card_creator(CardCreator(self._create_top_level_args()))
+            self._set_card_creator(CardCreator(self._create_top_level_args(flow)))
 
             current._update_env(
                 {"card": CardComponentCollector(self._logger, self.card_creator)}
@@ -223,6 +246,13 @@ class CardDecorator(StepDecorator):
             self.card_creator.create(mode="render", final=True, **create_options)
             self.card_creator.create(mode="refresh", final=True, **create_options)
 
+        # Unlink the config file if it exists
+        if self._config_file_name:
+            try:
+                os.unlink(self._config_file_name)
+            except Exception as e:
+                pass
+
     @staticmethod
     def _options(mapping):
         for k, v in mapping.items():
@@ -232,9 +262,13 @@ class CardDecorator(StepDecorator):
                 for value in v:
                     yield "--%s" % k
                     if not isinstance(value, bool):
-                        yield to_unicode(value)
+                        if isinstance(value, tuple):
+                            for val in value:
+                                yield to_unicode(val)
+                        else:
+                            yield to_unicode(value)
 
-    def _create_top_level_args(self):
+    def _create_top_level_args(self, flow):
         top_level_options = {
             "quiet": True,
             "metadata": self._metadata.TYPE,
@@ -247,4 +281,8 @@ class CardDecorator(StepDecorator):
             # We don't provide --with as all execution is taking place in
             # the context of the main process
         }
+        if self._config_values:
+            top_level_options["config-value"] = self._config_values
+            top_level_options["local-config-file"] = self._config_file_name
+
         return list(self._options(top_level_options))

metaflow/plugins/datatools/s3/s3op.py

@@ -722,8 +722,8 @@ def cli():
     pass
 
 
-@tracing.cli_entrypoint("s3op/list")
 @cli.command("list", help="List S3 objects")
+@tracing.cli_entrypoint("s3op/list")
 @click.option(
     "--recursive/--no-recursive",
     default=False,
@@ -782,8 +782,8 @@ def lst(
             print(format_result_line(idx, url.prefix, url.url, str(size)))
 
 
-@tracing.cli_entrypoint("s3op/put")
 @cli.command(help="Upload files to S3")
+@tracing.cli_entrypoint("s3op/put")
 @click.option(
     "--file",
     "files",
@@ -977,8 +977,8 @@ def _populate_prefixes(prefixes, inputs):
     return prefixes, is_transient_retry
 
 
-@tracing.cli_entrypoint("s3op/get")
 @cli.command(help="Download files from S3")
+@tracing.cli_entrypoint("s3op/get")
 @click.option(
     "--recursive/--no-recursive",
     default=False,

metaflow/plugins/kubernetes/kubernetes_cli.py

@@ -33,12 +33,12 @@ def kubernetes():
     pass
 
 
-@tracing.cli_entrypoint("kubernetes/step")
 @kubernetes.command(
     help="Execute a single task on Kubernetes. This command calls the top-level step "
     "command inside a Kubernetes pod with the given options. Typically you do not call "
     "this command directly; it is used internally by Metaflow."
 )
+@tracing.cli_entrypoint("kubernetes/step")
 @click.argument("step-name")
 @click.argument("code-package-sha")
 @click.argument("code-package-url")

metaflow/plugins/kubernetes/kubernetes_decorator.py

@@ -153,8 +153,8 @@ class KubernetesDecorator(StepDecorator):
     supports_conda_environment = True
     target_platform = "linux-64"
 
-    def __init__(self, attributes=None, statically_defined=False):
-        super(KubernetesDecorator, self).__init__(attributes, statically_defined)
+    def init(self):
+        super(KubernetesDecorator, self).init()
 
         if not self.attributes["namespace"]:
             self.attributes["namespace"] = KUBERNETES_NAMESPACE

metaflow/plugins/pypi/bootstrap.py

@@ -1,4 +1,5 @@
 import bz2
+import concurrent.futures
 import io
 import json
 import os
@@ -6,6 +7,9 @@ import shutil
 import subprocess
 import sys
 import tarfile
+import time
+
+import requests
 
 from metaflow.metaflow_config import DATASTORE_LOCAL_DIR
 from metaflow.plugins import DATASTORES
@@ -15,6 +19,18 @@ from . import MAGIC_FILE, _datastore_packageroot
 
 # Bootstraps a valid conda virtual environment composed of conda and pypi packages
 
+
+def timer(func):
+    def wrapper(*args, **kwargs):
+        start_time = time.time()
+        result = func(*args, **kwargs)
+        duration = time.time() - start_time
+        # print(f"Time taken for {func.__name__}: {duration:.2f} seconds")
+        return result
+
+    return wrapper
+
+
 if __name__ == "__main__":
     if len(sys.argv) != 5:
         print("Usage: bootstrap.py <flow_name> <id> <datastore_type> <architecture>")
@@ -47,6 +63,8 @@ if __name__ == "__main__":
 
     prefix = os.path.join(os.getcwd(), architecture, id_)
     pkgs_dir = os.path.join(os.getcwd(), ".pkgs")
+    conda_pkgs_dir = os.path.join(pkgs_dir, "conda")
+    pypi_pkgs_dir = os.path.join(pkgs_dir, "pypi")
     manifest_dir = os.path.join(os.getcwd(), DATASTORE_LOCAL_DIR, flow_name)
 
     datastores = [d for d in DATASTORES if d.TYPE == datastore_type]
@@ -64,77 +82,194 @@ if __name__ == "__main__":
         os.path.join(os.getcwd(), MAGIC_FILE),
         os.path.join(manifest_dir, MAGIC_FILE),
     )
-
     with open(os.path.join(manifest_dir, MAGIC_FILE)) as f:
         env = json.load(f)[id_][architecture]
 
-    # Download Conda packages.
-    conda_pkgs_dir = os.path.join(pkgs_dir, "conda")
-    with storage.load_bytes([package["path"] for package in env["conda"]]) as results:
-        for key, tmpfile, _ in results:
+    def run_cmd(cmd):
+        result = subprocess.run(
+            cmd, shell=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True
+        )
+        if result.returncode != 0:
+            print(f"Bootstrap failed while executing: {cmd}")
+            print("Stdout:", result.stdout)
+            print("Stderr:", result.stderr)
+            sys.exit(1)
+
+    @timer
+    def install_micromamba(architecture):
+        micromamba_dir = os.path.join(os.getcwd(), "micromamba")
+        micromamba_path = os.path.join(micromamba_dir, "bin", "micromamba")
+
+        if which("micromamba"):
+            return which("micromamba")
+        if os.path.exists(micromamba_path):
+            os.environ["PATH"] += os.pathsep + os.path.dirname(micromamba_path)
+            return micromamba_path
+
+        # Download and extract in one go
+        # TODO: Serve from cloudflare
+        url = f"https://micro.mamba.pm/api/micromamba/{architecture}/2.0.4"
+
+        # Prepare directory once
+        os.makedirs(os.path.dirname(micromamba_path), exist_ok=True)
+
+        # Stream and process directly to file
+        with requests.get(url, stream=True, timeout=30) as response:
+            if response.status_code != 200:
+                raise Exception(
+                    f"Failed to download micromamba: HTTP {response.status_code}"
+                )
+
+            decompressor = bz2.BZ2Decompressor()
+
+            # Process in memory without temporary files
+            tar_content = decompressor.decompress(response.raw.read())
+
+            with tarfile.open(fileobj=io.BytesIO(tar_content), mode="r:") as tar:
+                member = tar.getmember("bin/micromamba")
+                # Extract directly to final location
+                with open(micromamba_path, "wb") as f:
+                    f.write(tar.extractfile(member).read())
+
+        # Set executable permission
+        os.chmod(micromamba_path, 0o755)
+
+        # Update PATH only once at the end
+        os.environ["PATH"] += os.pathsep + os.path.dirname(micromamba_path)
+        return micromamba_path
+
+    @timer
+    def download_conda_packages(storage, packages, dest_dir):
+
+        def process_conda_package(args):
             # Ensure that conda packages go into architecture specific folders.
             # The path looks like REPO/CHANNEL/CONDA_SUBDIR/PACKAGE. We trick
             # Micromamba into believing that all packages are coming from a local
             # channel - the only hurdle is ensuring that packages are organised
             # properly.
-
-            # TODO: consider RAM disk
-            dest = os.path.join(conda_pkgs_dir, "/".join(key.split("/")[-2:]))
+            key, tmpfile, dest_dir = args
+            dest = os.path.join(dest_dir, "/".join(key.split("/")[-2:]))
             os.makedirs(os.path.dirname(dest), exist_ok=True)
             shutil.move(tmpfile, dest)
 
-    # Create Conda environment.
-    cmds = [
-        # TODO: check if mamba or conda are already available on the image
-        # TODO: micromamba installation can be pawned off to micromamba.py
-        f"""set -e;
-        if ! command -v micromamba >/dev/null 2>&1; then
-            mkdir -p micromamba;
-            python -c "import requests, bz2, sys; data = requests.get('https://micro.mamba.pm/api/micromamba/{architecture}/1.5.7').content; sys.stdout.buffer.write(bz2.decompress(data))" | tar -xv -C $(pwd)/micromamba bin/micromamba --strip-components 1;
+        os.makedirs(dest_dir, exist_ok=True)
+        with storage.load_bytes([package["path"] for package in packages]) as results:
+            with concurrent.futures.ThreadPoolExecutor() as executor:
+                executor.map(
+                    process_conda_package,
+                    [(key, tmpfile, dest_dir) for key, tmpfile, _ in results],
+                )
+            # for key, tmpfile, _ in results:
+
+            #     # TODO: consider RAM disk
+            #     dest = os.path.join(dest_dir, "/".join(key.split("/")[-2:]))
+            #     os.makedirs(os.path.dirname(dest), exist_ok=True)
+            #     shutil.move(tmpfile, dest)
+        return dest_dir
+
+    @timer
+    def download_pypi_packages(storage, packages, dest_dir):
+
+        def process_pypi_package(args):
+            key, tmpfile, dest_dir = args
+            dest = os.path.join(dest_dir, os.path.basename(key))
+            shutil.move(tmpfile, dest)
+
+        os.makedirs(dest_dir, exist_ok=True)
+        with storage.load_bytes([package["path"] for package in packages]) as results:
+            with concurrent.futures.ThreadPoolExecutor() as executor:
+                executor.map(
+                    process_pypi_package,
+                    [(key, tmpfile, dest_dir) for key, tmpfile, _ in results],
+                )
+            # for key, tmpfile, _ in results:
+            #     dest = os.path.join(dest_dir, os.path.basename(key))
+            #     shutil.move(tmpfile, dest)
+        return dest_dir
+
+    @timer
+    def create_conda_environment(prefix, conda_pkgs_dir):
+        cmd = f'''set -e;
+            tmpfile=$(mktemp);
+            echo "@EXPLICIT" > "$tmpfile";
+            ls -d {conda_pkgs_dir}/*/* >> "$tmpfile";
             export PATH=$PATH:$(pwd)/micromamba;
-            if ! command -v micromamba >/dev/null 2>&1; then
-                echo "Failed to install Micromamba!";
-                exit 1;
-            fi;
-        fi""",
-        # Create a conda environment through Micromamba.
-        f'''set -e;
-        tmpfile=$(mktemp);
-        echo "@EXPLICIT" > "$tmpfile";
-        ls -d {conda_pkgs_dir}/*/* >> "$tmpfile";
-        export PATH=$PATH:$(pwd)/micromamba;
-        export CONDA_PKGS_DIRS=$(pwd)/micromamba/pkgs;
-        micromamba create --yes --offline --no-deps --safety-checks=disabled --no-extra-safety-checks --prefix {prefix} --file "$tmpfile";
-        rm "$tmpfile"''',
-    ]
-
-    # Download PyPI packages.
-    if "pypi" in env:
-        pypi_pkgs_dir = os.path.join(pkgs_dir, "pypi")
-        with storage.load_bytes(
-            [package["path"] for package in env["pypi"]]
-        ) as results:
-            for key, tmpfile, _ in results:
-                dest = os.path.join(pypi_pkgs_dir, os.path.basename(key))
-                os.makedirs(os.path.dirname(dest), exist_ok=True)
-                shutil.move(tmpfile, dest)
-
-        # Install PyPI packages.
-        cmds.extend(
-            [
-                f"""set -e;
-                export PATH=$PATH:$(pwd)/micromamba;
-                export CONDA_PKGS_DIRS=$(pwd)/micromamba/pkgs;
-                micromamba run --prefix {prefix} python -m pip --disable-pip-version-check install --root-user-action=ignore --no-compile {pypi_pkgs_dir}/*.whl --no-user"""
-            ]
-        )
+            export CONDA_PKGS_DIRS=$(pwd)/micromamba/pkgs;
+            export MAMBA_NO_LOW_SPEED_LIMIT=1;
+            export MAMBA_USE_INDEX_CACHE=1;
+            export MAMBA_NO_PROGRESS_BARS=1;
+            export CONDA_FETCH_THREADS=1;
+            micromamba create --yes --offline --no-deps \
+                --safety-checks=disabled --no-extra-safety-checks \
+                --prefix {prefix} --file "$tmpfile" \
+                --no-pyc --no-rc --always-copy;
+            rm "$tmpfile"'''
+        run_cmd(cmd)
 
-    for cmd in cmds:
-        result = subprocess.run(
-            cmd, shell=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE
-        )
-        if result.returncode != 0:
-            print(f"Bootstrap failed while executing: {cmd}")
-            print("Stdout:", result.stdout.decode())
-            print("Stderr:", result.stderr.decode())
-            sys.exit(1)
+    @timer
+    def install_pypi_packages(prefix, pypi_pkgs_dir):
+
+        cmd = f"""set -e;
+            export PATH=$PATH:$(pwd)/micromamba;
+            export CONDA_PKGS_DIRS=$(pwd)/micromamba/pkgs;
+            micromamba run --prefix {prefix} python -m pip --disable-pip-version-check \
+                install --root-user-action=ignore --no-compile --no-index \
+                --no-cache-dir --no-deps --prefer-binary \
+                --find-links={pypi_pkgs_dir}  --no-user \
+                --no-warn-script-location --no-input \
+                {pypi_pkgs_dir}/*.whl
+            """
+        run_cmd(cmd)
+
+    @timer
+    def setup_environment(
+        architecture, storage, env, prefix, conda_pkgs_dir, pypi_pkgs_dir
+    ):
+        with concurrent.futures.ThreadPoolExecutor(max_workers=10) as executor:
+            # install micromamba, download conda and pypi packages in parallel
+            futures = {
+                "micromamba": executor.submit(install_micromamba, architecture),
+                "conda_pkgs": executor.submit(
+                    download_conda_packages, storage, env["conda"], conda_pkgs_dir
+                ),
+            }
+            if "pypi" in env:
+                futures["pypi_pkgs"] = executor.submit(
+                    download_pypi_packages, storage, env["pypi"], pypi_pkgs_dir
+                )
+
+            # create conda environment after micromamba is installed and conda packages are downloaded
+            done, _ = concurrent.futures.wait(
+                [futures["micromamba"], futures["conda_pkgs"]],
+                return_when=concurrent.futures.ALL_COMPLETED,
+            )
+
+            for future in done:
+                future.result()
+
+            # start conda environment creation
+            futures["conda_env"] = executor.submit(
+                create_conda_environment, prefix, conda_pkgs_dir
+            )
+
+            if "pypi" in env:
+                # install pypi packages after conda environment is created and pypi packages are downloaded
+                done, _ = concurrent.futures.wait(
+                    [futures["conda_env"], futures["pypi_pkgs"]],
+                    return_when=concurrent.futures.ALL_COMPLETED,
+                )
+
+                for future in done:
+                    future.result()
+
+                # install pypi packages
+                futures["pypi_install"] = executor.submit(
+                    install_pypi_packages, prefix, pypi_pkgs_dir
+                )
+                # wait for pypi packages to be installed
+                futures["pypi_install"].result()
+            else:
+                # wait for conda environment to be created
+                futures["conda_env"].result()
+
+    setup_environment(architecture, storage, env, prefix, conda_pkgs_dir, pypi_pkgs_dir)