services

#!/usr/bin/env python3


import argparse
import copy
import os
import subprocess
from typing import Dict, List, Optional
import yaml


ROOT = os.environ.get("ROOT", os.getcwd())


# If any file in these folders are changed, rebuild all services
REBUILD_FOLDERS = [
    "shared"
]


# If any file in these service folders are changed, ignore it.
# If the folder does not have a `metadata.yaml` file, the folder is ignored
# anyway.
IGNORE_FOLDERS = []


def get_args():
    """
    Retrieve arguments from the commandline
    """

    parser = argparse.ArgumentParser()
    parser.add_argument("--env-only", default=False, action="store_true")
    parser.add_argument("--graph", default=False, action="store_true")
    parser.add_argument("--reverse", default=False, action="store_true")
    parser.add_argument("--exclude", action="append")
    parser.add_argument("--deps-of", action="append")
    parser.add_argument("--changed-since", default="0")

    return parser.parse_args()


def get_metadata(service_name: str) -> dict:
    """
    Get metadata for a service
    """

    metafile = os.path.join(ROOT, service_name, "metadata.yaml")
    if not os.path.isfile(metafile):
        raise ValueError("Metadata file not found for {}".format(service_name))

    with open(metafile) as fp:
        metadata = yaml.load(fp, Loader=yaml.SafeLoader)

    return metadata


def get_services(
        env_only: bool = False,
        deps_of: Optional[List[str]] = None,
        exclude: Optional[List[str]] = None
    ) -> Dict[str, dict]:
    """
    Return the list of services

    env_only: only return services that support environments (after parsing dependencies)
    deps_of: only return services that are a dependency of another service (including itself)
    exclude: exclude these services from the list (after parsing dependencies)
    """

    # Gather all services
    if deps_of is None:
        services = {
            # Mapping name: reverse dependencies
            name: {"deps": [], "rdeps": [], "metadata": get_metadata(name)}
            for name in os.listdir(ROOT)
            if os.path.isfile(os.path.join(ROOT, name, "metadata.yaml"))
        }
    else:
        services = {}
        to_scan = deps_of
        while len(to_scan) > 0:
            name = to_scan.pop(0)
            services[name] = {"deps": [], "rdeps": [], "metadata": get_metadata(name)}
            for dependency in services[name]["metadata"].get("dependencies", []):
                if dependency not in services and dependency not in to_scan:
                    to_scan.append(dependency)

    # If needed, remove services which don't support environments
    if env_only:
        for name in list(services.keys()):
            if not services[name]["metadata"].get("flags", {}).get("environment", True):
                del services[name]

    # Filter out excluded names
    return {k: v for k, v in services.items() if k not in IGNORE_FOLDERS + (exclude or [])}


def make_service_graph(
        services: Dict[str, dict],
        changed_since: str = "0"
    ) -> List[str]:
    """
    Retrieve services in dependency order
    """

    to_scan = []

    # Parse dependencies
    for name in services.keys():
        if len(services[name]["metadata"].get("dependencies", [])) == 0:
            to_scan.append(name)
        for dep in services[name]["metadata"].get("dependencies", []):
            if name == dep:
                continue
            # This service depends on all the services
            elif dep == "*":
                for dname in services.keys():
                    # Only inject if this is not the same service
                    if dname != name:
                        services[name]["deps"].append(dname)
                        services[dname]["rdeps"].append(name)
            elif dep not in services:
                raise ValueError("Dependency {} of {} not found".format(dep, name))
            else:
                services[name]["deps"].append(dep)
                # Inject as a reverse dependency
                services[dep]["rdeps"].append(name)

    # Scan all services until there are no more services to scan.
    count = 0
    retval = []
    while len(to_scan) > 0:
        retval_line = copy.deepcopy(to_scan)
        to_scan = []
        for name in retval_line:
            for rname in services[name]["rdeps"]:
                services[rname]["deps"].remove(name)
                if len(services[rname]["deps"]) == 0:
                    to_scan.append(rname)
        count += len(retval_line)
        retval.append(retval_line)


    # If there is a discrepancy, this means that there is a circular
    # dependency.
    if count != len(services):
        for key, value in services.items():
            print(key, value)
        raise ValueError("Potential circular dependency found: mismatch between number of services: {} vs {}".format(len(retval), len(services)))

    # If `--changed-since` is absent or does not have a valid commit ID, return the list of services
    if changed_since == "0":
        return retval

    # If `--changed-since` has a valid commit ID, we need to check which 
    process = subprocess.run(["git", "diff", changed_since, "--name-only"], capture_output=True, text=True)
    process.check_returncode()

    changed = []
    for filename in process.stdout.split("\n"):
        basename = filename.split("/", 1)[0]

        # If one file from the special folders has changed, all services need
        # to be rebuilt.
        if basename in REBUILD_FOLDERS:
            return retval

        if basename in services:
            changed.append(basename)

    updated_services = copy.deepcopy(retval)
    for index, service_line in enumerate(retval):
        for service in service_line:
            if service not in changed:
                updated_services[index].remove(service)

    # Only return each service once
    return [service_line for service_line in updated_services if service_line]


if __name__ == "__main__":
    args = get_args()
    services = make_service_graph(
        get_services(args.env_only, args.deps_of, args.exclude),
        args.changed_since
    )
    if args.reverse:
        services = services[::-1]

    for service_line in services:
        if args.graph:
            print(",".join(service_line))
        else:
            for service in service_line:
                print(service)