eo4eu-data-utils 0.15.0

General data access utilities for EO4EU

Data utilities for EO4EU

This package provides classes and functions that help with common tasks involving:

• Reading data from configMaps and secrets
• Uploading to and downloading from S3 buckets
• Configuring components in general

Installation

eo4eu-data-utils is published on PyPI and can be installed with pip anywhere. You can look for the latest version and pin that in your requirements.txt or what-have-you.

Usage

For example usage of this package, you may refer to post-pro, jupyter-openfaas or jupyter-proxy.

ConfigMaps and Secrets

You can use the legacy API if you want something similar to get_configMap and get_secret:

from eo4eu_data_utils.legacy import ClusterAccess

access = ClusterAccess()

BOTO_CONFIG = {
    "region_name":           access.cfgmap("s3-access", "region_name"),
    "endpoint_url":          access.cfgmap("s3-access", "endpoint_url"),
    "aws_access_key_id":     access.secret("s3-access-scr", "aws_access_key_id"),
    "aws_secret_access_key": access.secret("s3-access-scr", "aws_secret_access_key"),
}

It is recommended that you use eo4eu_data_utils.config.Config instead, which is described after S3 Access.

S3 Access

The old StorageUtils class is supported:

from eo4eu_data_utils.legacy import StorageUtils

ap = StorageUtils(config_boto=CONFIG_BOTO, config_cloudpath=CONFIG_CLOUD)

But the new one is more convenient; you specify the bucket when creating it and you don't have to keep track of it for every call:

from eo4eu_data_utils.storage import Datastore

datastore = Datastore(
    config = CONFIG_BOTO,  # exactly the same config dict as before
    bucket = BUCKET_NAME
)

List files in bucket:

files = datastore.list_files(subfolder, subsubfolder, ...)  # paths for (nested) subfolders are optional

Download and upload bytes:

file_data = datastore.download(s3_key)
succeeded: bool = datastore.upload(other_s3_key, file_data)

Download and upload files:

dl_succeeded: bool = datastore.download_to(s3_key, local_path)
up_succeeded: bool = datastore.upload_from(local_path, s3_key)

Download and upload many files at once:

dl_s3_keys = ["data/d0.csv", "data/d1.csv", "data/img.tiff"]
dl_output_dir = "download_dir"
dl_result = datastore.download_many(dl_s3_keys, dl_output_dir)

up_s3_keys = ["output/result_0/meta.json", "output/result_0/r0.csv"]
up_input_dir = "output"
up_result = datastore.upload_many(up_s3_keys, up_input_dir)

Here the variables dl_result and up_result are of the class eo4eu_data_utils.storage.TransferResult and contain the succeeded and failed transfers. Example:

for success in dl_result.succeeded:
    logger.info(f"Downloaded {success.src} to {success.dst}")

for failure in dl_result.failed:
    logger.warning(f"Failed to download {failure.src} to {failure.dst}")

You can check the number of successses/failures by way of TransferResult.succeeded_num and TransferResult.failed_num, or just get the len of the above lists.

Configuration

The Config class allows you to define a configuration dict and fill it in different ways depending on whether you're on dev or prod. For example:

from eo4eu_data_utils.config import Config, Try

unfilled_config = Config(
    boto = {
        "region_name":           Try.cfgmap("s3-access", "region_name"),
        "endpoint_url":          Try.cfgmap("s3-access", "endpoint_url"),
        "aws_access_key_id":     Try.secret("s3-access-scr", "aws_access_key_id"),
        "aws_secret_access_key": Try.secret("s3-access-scr", "aws_secret_access_key"),
    },
    eo4eu = {
        "namespace":      Try.cfgmap("eo4eu", "namespace"),
        "s3_bucket_name": Try.cfgmap("eo4eu", "s3-bucket-name"),
    },
    # ...
)

The values may be accessed as nested attributes or dict items:

# all of these are valid
key_id = config.boto.aws_access_key_id
key_id = config.boto["aws_access_key_id"]
key_id = config["boto"].aws_access_key_id
key_id = config["boto"]["aws_access_key_id"]

This means that config.boto is not a python dict. If you need a dict, you can convert it like so:

client = boto3.client(**config.boto.to_dict())

Filling the configuration

On prod, you can fill an unfilled config from the configMaps and secrets on the cluster:

config = unfilled_config.use_files().fill()

On dev, you can fill it from environment variables:

config = unfilled_config.use_env().fill()

For the previous example, the environment variables must be of the form:

export CONFIGMAPS_S3_ACCESS_REGION_NAME=
export CONFIGMAPS_S3_ACCESS_ENDPOINT_URL=
export CONFIGMAPS_EO4EU_NAMESPACE=
export CONFIGMAPS_EO4EU_S3_BUCKET_NAME=

export SECRETS_S3_ACCESS_SCR_AWS_ACCESS_KEY_ID=
export SECRETS_S3_ACCESS_SCR_AWS_SECRET_ACCESS_KEY=

Some common configs are defined in eo4eu_data_utils.config, you may want to take a look at them.

You can automatically convert the Try values to different types and set default values to be used in case they are not found:

from eo4eu_data_utils.config import Config, Try

unfilled_config = Config(
    # ...
    elasticsearch = {
        "username":    Try.secret("els-access-scr", "username"),
        "password":    Try.secret("els-access-scr", "password"),
        "local_dir":   Try.cfgmap("els", "local").default("els_dir").to_path(),  # i.e. pathlib.Path
        "max_retries": Try.cfgmap("els", "max_retries").default(10).to_int()
    }
)

In the above example, it should be noted that the order of methods matters. If /configmaps/els/local doesn't exist, it will be replaced with the default value "els_dir" and then converted to a path. If the order was switched, it would first try to convert the missing value to a path, fail, and then return the string "els_dir". It is recommended that you put all your type converters at the end to ensure the result is of the desired type.

A third constructor for Try is provided: Try.source. This is different to cfgmap and secret in that it doesn't assume it's a file in /configmaps or /secrets. It's there to provide extra configuration, for example:

from eo4eu_data_utils.config import Config, Try

unfilled_config = Config(
    # ...
    testing = {
        "raise_exceptions": Try.source("debug", "raise_exceptions").default(False).to_bool(),
        "max_retries":      Try.source("http", "max_retries").default(5).to_int(),
    },
)

These can be filled from environment variables:

config = unfilled_config.use_env().fill()

With:

export DEBUG_RAISE_EXCEPTIONS=true
export HTTP_MAX_RETRIES=10

Or, alternatively, you can use a simple python dictionary:

input = {
    "debug": {
        "raise_exceptions": True,
        "delete_output": False,  # extra keys get ignored
    },
    "http": {
        "max_retries": 10
    },
}

config = unfilled_config.use_dict(input).fill()

Or a JSON string:

input ="""{
    "debug": {
        "raise_exceptions": true
    },
    "http": {
        "max_retries": 10
    }
}"""

config = unfilled_config.use_json(input).fill()

You can chain different sources of inputs, and the config will look through them until it finds a match.

config = unfilled_config.use_env().use_dict(input).fill()

Finally, passing raise_on_err = True into Config.fill will raise a ValueError if a field was unable to be filled:

try:
    config = unfilled_config.use_files().fill(raise_on_err = True)
except Exception as e:
    logger.error(f"Unable to fill configuration: {e}")