rs_server_catalog/data_management/user_handler.md

<< Back to index

This library contains all functions needed for the fastAPI middleware.

adapt_links(content, object_name, current_user='', current_collection_id='')

Adapt all the links that are outside from the collection section with the given user and collection name, then the ones inside with the user and collection names they already contain.

Parameters:

Name	Type	Description	Default
content	dict	The response content from the middleware	required
object_name	str	Type of object we want to also update.	required
current_user	str	The user id that is currently connected.	''
current_collection_id	str	The collection ID used to rewrite links outside the object list.	''

Returns:

Name	Type	Description
dict	dict	The content passed in parameter with adapted links

Source code in docs/rs-server/services/catalog/rs_server_catalog/data_management/user_handler.py

def adapt_links(content: dict, object_name: str, current_user: str = "", current_collection_id: str = "") -> dict:
    """Adapt all the links that are outside from the collection section with the given user and collection name,
    then the ones inside with the user and collection names they already contain.

    Args:
        content (dict): The response content from the middleware
        'call_next' loaded in json format.
        object_name (str): Type of object we want to also update.
        current_user (str): The user id that is currently connected.
        current_collection_id (str): The collection ID used to rewrite links outside the object list.

    Returns:
        dict: The content passed in parameter with adapted links
    """
    # Adapt links outside of objects with current user/collection situation
    links = content["links"]
    for link in links:
        link_parser = urlparse(link["href"])
        new_path = add_user_prefix(link_parser.path, current_user, current_collection_id)
        link["href"] = link_parser._replace(path=new_path).geturl()

    # Go through each object and apply corrections to the links using the object's info
    for i in range(len(content[object_name])):
        content[object_name][i] = adapt_object_links(content[object_name][i], current_user)

    return content

adapt_object_links(object_content, current_user='')

Adapt all the links from a collection using the user and collection name they already contain, so the user can access them correctly

Parameters:

Name	Type	Description	Default
object_content	dict	The STAC object (item or collection) whose links need to be adapted.	required
current_user	str	The currently connected user identifier (used as fallback owner).	''

Returns: dict: The collection passed in parameter with adapted links

Source code in docs/rs-server/services/catalog/rs_server_catalog/data_management/user_handler.py

def adapt_object_links(object_content: dict, current_user: str = "") -> dict:
    """Adapt all the links from a collection using the user and collection name they already contain,
    so the user can access them correctly

    Args:
        object_content (dict): The STAC object (item or collection) whose links need to be adapted.
        current_user (str): The currently connected user identifier (used as fallback owner).
    Returns:
        dict: The collection passed in parameter with adapted links
    """
    user = collection_id = feature_id = ""

    # Case when object is an item
    if "properties" in object_content and "collection" in object_content:
        object_content, user = remove_owner_from_collection_name_in_feature(object_content, current_user)
        collection_id = object_content["collection"]
        feature_id = object_content["id"]

    # Case when object is a collection
    elif "id" in object_content:
        object_content, user = remove_owner_from_collection_name_in_collection(object_content, current_user)
        collection_id = object_content["id"]

    # Update links with user, collection and feature values retrieved from previous steps
    links = object_content.get("links", [])
    for j, link in enumerate(links):
        link_parser = urlparse(link["href"])
        new_path = add_user_prefix(link_parser.path, user, collection_id, feature_id)
        links[j]["href"] = link_parser._replace(path=new_path).geturl()

    return object_content

add_user_prefix(path, user, collection_id, feature_id='')

Modify the RS server backend catalog endpoint to get the RS server frontend endpoint

Parameters:

Name	Type	Description	Default
path	str	RS server backend endpoint.	required
user	str	The user ID.	required
collection_id	str	The collection id.	required
feature_id	str	The feature id.	''

Returns:

Name	Type	Description
str	str	The RS server frontend endpoint.

Source code in docs/rs-server/services/catalog/rs_server_catalog/data_management/user_handler.py

def add_user_prefix(
    path: str,
    user: str | None,
    collection_id: str | None,
    feature_id: str = "",
) -> str:
    """
    Modify the RS server backend catalog endpoint to get the RS server frontend endpoint

    Args:
        path (str): RS server backend endpoint.
        user (str): The user ID.
        collection_id (str): The collection id.
        feature_id (str): The feature id.

    Returns:
        str: The RS server frontend endpoint.
    """
    new_path = path

    if path == "/collections":
        new_path = CATALOG_COLLECTIONS

    elif path == "/search":
        new_path = CATALOG_SEARCH

    elif user and (path == "/"):
        new_path = CATALOG_PREFIX + "/"

    elif user and collection_id and (path == CATALOG_COLLECTIONS + f"/{user}_{collection_id}"):
        new_path = CATALOG_COLLECTIONS + f"/{user}:{collection_id}"

    elif user and collection_id and (path == CATALOG_COLLECTIONS + f"/{user}_{collection_id}/items"):
        new_path = CATALOG_COLLECTIONS + f"/{user}:{collection_id}/items"

    elif user and collection_id and (path == CATALOG_COLLECTIONS + f"/{user}_{collection_id}/queryables"):
        new_path = CATALOG_COLLECTIONS + f"/{user}:{collection_id}/queryables"

    elif (
        user
        and collection_id
        and (f"/collections/{user}_{collection_id}/items" in path or f"/collections/{collection_id}/items" in path)
    ):  # /catalog/.../items/item_id
        new_path = CATALOG_COLLECTIONS + f"/{user}:{collection_id}/items/{feature_id}"

    return new_path

collection_id_without_owner_id(collection_id, owner_id)

Returns collection_id without owner_id prefix

Source code in docs/rs-server/services/catalog/rs_server_catalog/data_management/user_handler.py

def collection_id_without_owner_id(collection_id: str, owner_id: str) -> str:
    """Returns collection_id without owner_id prefix"""
    return collection_id.removeprefix(f"{owner_id}_")

get_user(endpoint_user, apikey_user)

Retrieve the user identifier based on provided parameters. Default is the current running user (used for local mode in general)

Parameters:

Name	Type	Description	Default
endpoint_user	str	User identifier from the endpoint.	required
apikey_user	str	User identifier from the API key.	required

Returns:

Name	Type	Description
str		The user identifier.

Source code in docs/rs-server/services/catalog/rs_server_catalog/data_management/user_handler.py

def get_user(endpoint_user: str | None, apikey_user: str | None):
    """Retrieve the user identifier based on provided parameters. Default is the
    current running user (used for local mode in general)

    Args:
        endpoint_user (str): User identifier from the endpoint.
        apikey_user (str): User identifier from the API key.

    Returns:
        str: The user identifier.
    """
    if endpoint_user:
        return endpoint_user
    if apikey_user:
        return apikey_user
    return os.getenv("RSPY_HOST_USER", default=getpass.getuser())

owner_id_and_collection_id(owner_id, collection_id)

Returns collection_id with owner_id prefix

Source code in docs/rs-server/services/catalog/rs_server_catalog/data_management/user_handler.py

def owner_id_and_collection_id(owner_id: str, collection_id: str) -> str:
    """Returns collection_id with owner_id prefix"""
    return collection_id if collection_id.startswith(f"{owner_id}_") else f"{owner_id}_{collection_id}"

remove_owner_from_collection_name_in_collection(collection, current_user='')

Remove the owner name from the given collection name. The owner name used is the "owner" field of the collection if there is any, or by default the currently connected user. Returns the updated collection and the owner name actually removed. If nothing was removed, returns the original collection and an empty owner name.

Parameters:

Name	Type	Description	Default
collection	dict	A dictionary that contains metadata about the collection content like the id of the collection.	required
current_user	str	current user connected (optional)	''

Returns:

Name	Type	Description
dict	dict	The collection without the owner name in the id section.
str	str	the name removed, if any.

Source code in docs/rs-server/services/catalog/rs_server_catalog/data_management/user_handler.py

def remove_owner_from_collection_name_in_collection(collection: dict, current_user: str = "") -> tuple[dict, str]:
    """Remove the owner name from the given collection name.
    The owner name used is the "owner" field of the collection if there is any,
    or by default the currently connected user.
    Returns the updated collection and the owner name actually removed.
    If nothing was removed, returns the original collection and an empty owner name.

    Args:
        collection (dict): A dictionary that contains metadata
            about the collection content like the id of the collection.
        current_user (str): current user connected (optional)

    Returns:
        dict: The collection without the owner name in the id section.
        str: the name removed, if any.
    """
    if "owner" in collection:
        user = collection["owner"]
    else:
        user = current_user

    if collection["id"].startswith(f"{user}_"):
        collection["id"] = collection["id"].removeprefix(f"{user}_")
        return collection, user
    return collection, ""

remove_owner_from_collection_name_in_feature(feature, current_user='')

Remove the owner name from the collection name in the feature. The owner name used is the "owner" field of the properties if there is any, or by default the currently connected user. Returns the updated feature and the owner name actually removed. If nothing was removed, returns the original feature and an empty owner name.

Parameters:

Name	Type	Description	Default
feature	dict	a geojson that contains georeferenced data and metadata like the collection name.	required
current_user	str	current user connected (optional)	''

Returns:

Name	Type	Description
dict	dict	the feature with a new collection name without the owner name.
str	str	the name removed, if any.

Source code in docs/rs-server/services/catalog/rs_server_catalog/data_management/user_handler.py

def remove_owner_from_collection_name_in_feature(feature: dict, current_user: str = "") -> tuple[dict, str]:
    """Remove the owner name from the collection name in the feature.
    The owner name used is the "owner" field of the properties if there is any,
    or by default the currently connected user.
    Returns the updated feature and the owner name actually removed.
    If nothing was removed, returns the original feature and an empty owner name.

    Args:
        feature (dict): a geojson that contains georeferenced
            data and metadata like the collection name.
        current_user (str): current user connected (optional)

    Returns:
        dict: the feature with a new collection name without the owner name.
        str: the name removed, if any.
    """
    if "owner" in feature["properties"]:
        user = feature["properties"]["owner"]
    else:
        user = current_user

    if feature["collection"].startswith(f"{user}_"):
        feature["collection"] = feature["collection"].removeprefix(f"{user}_")
        return feature, user
    return feature, ""

reroute_url(request, ids_dict)

Get the RS Server backend catalog endpoints.

Parameters:

Name	Type	Description	Default
request	Request	The client request	required
ids_dict	str	Dictionary to easily access important data of our request	required

Raises:

Type	Description
ValueError	If the path is not valid.

Source code in docs/rs-server/services/catalog/rs_server_catalog/data_management/user_handler.py

def reroute_url(  # type: ignore # pylint: disable=too-many-branches,too-many-statements
    request: Request,
    ids_dict: dict[str, Any],
):
    """Get the RS Server backend catalog endpoints.

    Args:
        request (Request): The client request
        ids_dict (str): Dictionary to easily access important data of our request

    Raises:
        ValueError: If the path is not valid.
    """
    path = request.scope["path"]
    method = request.method
    patterns = [r"/favicon.ico"]

    # Catch one endpoint of the following list
    regexp_list = [
        "/",
        CATALOG_PREFIX,
        CATALOG_PREFIX + "/",
        CATALOG_SEARCH,
        CATALOG_PREFIX + "/queryables",
        CATALOG_PREFIX + "/api",
        CATALOG_PREFIX + "/api.html",
        CATALOG_PREFIX + "/docs/oauth2-redirect",
        CATALOG_PREFIX + "/conformance",
    ]
    if any(re.fullmatch(pattern, path) for pattern in regexp_list):
        # Don't validate other conditions if we alredy matched the previous regexps
        pass

    # Catch authentication endpoints (path should be left as it is in this case)
    elif path.startswith(f"{oauth2.AUTH_PREFIX}/"):
        pass

    # The endpoint PUT "/catalog/collections" does not exists.
    elif path.rstrip("/") == CATALOG_COLLECTIONS and method != "PUT":
        path = CATALOG_COLLECTIONS

    # Catch endpoint /catalog/collections/[{owner_id}:]{collection_id}/bulk_items
    elif match := re.fullmatch(BULK_ITEMS_REGEX, path):
        groups = match.groupdict()
        ids_dict["owner_id"] = get_user(groups["owner_id"], ids_dict["user_login"])
        ids_dict["collection_ids"].append(collection_id_without_owner_id(groups["collection_id"], ids_dict["owner_id"]))
        path = (
            CATALOG_COLLECTIONS
            + f"/{owner_id_and_collection_id(ids_dict['owner_id'], ids_dict['collection_ids'][0])}/bulk_items"
        )

    # Catch endpoint /catalog/collections/[{owner_id}:]{collection_id}/queryables
    elif match := re.fullmatch(COLLECTIONS_QUERYABLES_REGEX, path):
        groups = match.groupdict()
        ids_dict["owner_id"] = get_user(groups["owner_id"], ids_dict["user_login"])
        ids_dict["collection_ids"].append(collection_id_without_owner_id(groups["collection_id"], ids_dict["owner_id"]))
        path = (
            CATALOG_COLLECTIONS
            + f"/{owner_id_and_collection_id(ids_dict['owner_id'], ids_dict['collection_ids'][0])}/queryables"
        )

    # Catch all other endpoints.
    elif match := re.match(CATALOG_OWNER_ID_STAC_ENDPOINT_REGEX, path):
        groups = match.groupdict()
        if groups["owner_collection_id"]:
            # protection for more than one : (example-> /catalog/collections/ownerId:collection:Id/items)
            # the list owner_collection_id_split has one or at most two members (note the maxsplit = 1)
            owner_collection_id_split = groups["owner_collection_id"].lstrip("/").split(":", 1)
            if len(owner_collection_id_split) == 1:
                # the following handles the absence of the ownerId param, for endpoints like:
                # /catalog/collections/collectionId/items
                ids_dict["owner_id"] = get_user(None, ids_dict["user_login"])
                ids_dict["collection_ids"].append(
                    collection_id_without_owner_id(owner_collection_id_split[0], ids_dict["owner_id"]),
                )
            else:
                # the following handles the presence of the ownerId param, for endpoints like:
                # /catalog/collections/ownerId:collectionId/items
                ids_dict["owner_id"] = owner_collection_id_split[0]
                ids_dict["collection_ids"].append(
                    collection_id_without_owner_id(owner_collection_id_split[1], ids_dict["owner_id"]),
                )

        # /catalog/collections/{owner_id}:{collection_id}
        # case is the same for PUT / POST / DELETE, but needs different paths
        collection_id = owner_id_and_collection_id(ids_dict["owner_id"], ids_dict["collection_ids"][0])
        if groups["items"] is None and method != "DELETE":
            path = CATALOG_COLLECTIONS + f"/{collection_id}"
        else:
            ids_dict["item_id"] = groups["item_id"]
            if ids_dict["item_id"] is None:
                if "/items" in path:
                    path = CATALOG_COLLECTIONS + f"/{collection_id}/items"
                else:
                    path = CATALOG_COLLECTIONS + f"/{collection_id}"
            else:
                ids_dict["item_id"] = ids_dict["item_id"][1:]
                path = CATALOG_COLLECTIONS + f"/{collection_id}/items/{ids_dict['item_id']}"

    elif not any(re.fullmatch(pattern, path) for pattern in patterns):
        path = ""
    # Finally, update the path of the request with the new route
    if path != request.scope["path"]:
        logger.debug(f"Rerouting {request.scope['path']} => {path}")
        request.scope["path"] = path