Unverified Commit 21e7a222 authored by Marius van den Beek's avatar Marius van den Beek Committed by GitHub
Browse files

Merge pull request #18494 from mvdbeek/fix_invalid_id_error_message 

[24.1] Increase API robustness to invalid requests, improve compressed data serving
parents 545ef625 92bfaa63

lib/galaxy/datatypes/data.py

+22 −20
Original line number Diff line number Diff line
@@ -484,11 +484,12 @@ class Data(metaclass=DataMeta): 


    def _serve_binary_file_contents_as_text(self, trans, data, headers, file_size, max_peek_size):

        headers["content-type"] = "text/html"

        with open(data.get_file_name(), "rb") as fh:

            return (

                trans.fill_template_mako(

                    "/dataset/binary_file.mako",

                    data=data,

                file_contents=open(data.get_file_name(), "rb").read(max_peek_size),

                    file_contents=fh.read(max_peek_size),

                    file_size=util.nice_size(file_size),

                    truncated=file_size > max_peek_size,

                ),
@@ -502,12 +503,13 @@ class Data(metaclass=DataMeta): 
        if not preview or isinstance(data.datatype, images.Image) or file_size < max_peek_size:

            return self._yield_user_file_content(trans, data, data.get_file_name(), headers), headers



        with compression_utils.get_fileobj(data.get_file_name(), "rb") as fh:

            # preview large text file

            headers["content-type"] = "text/html"

            return (

                trans.fill_template_mako(

                    "/dataset/large_file.mako",

                truncated_data=open(data.get_file_name(), "rb").read(max_peek_size),

                    truncated_data=fh.read(max_peek_size),

                    data=data,

                ),

                headers,

lib/galaxy/datatypes/tabular.py

+9 −8
Original line number Diff line number Diff line
@@ -193,10 +193,11 @@ class TabularData(Text): 
                return open(dataset.get_file_name(), mode="rb"), headers

            else:

                headers["content-type"] = "text/html"

                with compression_utils.get_fileobj(dataset.get_file_name(), "rb") as fh:

                    return (

                        trans.fill_template_mako(

                            "/dataset/large_file.mako",

                        truncated_data=open(dataset.get_file_name()).read(max_peek_size),

                            truncated_data=fh.read(max_peek_size),

                            data=dataset,

                        ),

                        headers,

lib/galaxy/managers/hdas.py

+4 −1
Original line number Diff line number Diff line
@@ -311,7 +311,10 @@ class HDAManager( 


        truncated = preview and os.stat(file_path).st_size > MAX_PEEK_SIZE

        with get_fileobj(file_path) as fh:

            try:

                hda_data = fh.read(MAX_PEEK_SIZE)

            except UnicodeDecodeError:

                raise exceptions.RequestParameterInvalidException("Cannot generate text preview for dataset.")

        return truncated, hda_data



    # .... annotatable

lib/galaxy/tools/parameters/basic.py

+2 −1
Original line number Diff line number Diff line
@@ -2053,7 +2053,8 @@ def src_id_to_item( 
        item = sa_session.get(src_to_class[value["src"]], decoded_id)

    except KeyError:

        raise ValueError(f"Unknown input source {value['src']} passed to job submission API.")

    assert item

    if not item:

        raise ValueError("Invalid input id passed to job submission API.")

    item.extra_params = {k: v for k, v in value.items() if k not in ("src", "id")}

    return item

lib/galaxy/web/framework/middleware/statsd.py

+7 −2
Original line number Diff line number Diff line
@@ -28,8 +28,13 @@ class StatsdMiddleware: 
        start_time = time.time()

        req = self.application(environ, start_response)

        dt = int((time.time() - start_time) * 1000)

        page = environ.get("controller_action_key", None) or environ.get("PATH_INFO", "NOPATH").strip("/").replace(

            "/", "."

        page = (

            environ.get("controller_action_key", None)

            or environ.get("PATH_INFO", "NOPATH")

            .strip("/")

            .replace("/", ".")

            .encode("ascii", errors="replace")

            .decode()

        )

        self.galaxy_stasd_client.timing(page, dt)

        try: