json: Replaced json results processor with a more comprehensive one

wlauto/core/configuration.py

@@ -243,6 +243,13 @@ class RebootPolicy(object):
         else:
             return cmp(self.policy, other)
 
+    def to_pod(self):
+        return self.policy
+
+    @staticmethod
+    def from_pod(pod):
+        return RebootPolicy(pod)
+
 
 class RunConfigurationItem(object):
     """

wlauto/exceptions.py

@@ -141,3 +141,20 @@ class WorkerThreadError(WAError):
         message = 'Exception of type {} occured on thread {}:\n'.format(orig_name, thread)
         message += '{}\n{}: {}'.format(get_traceback(self.exc_info), orig_name, orig)
         super(WorkerThreadError, self).__init__(message)
+
+
+class SerializerSyntaxError(Exception):
+    """
+    Error loading a serialized structure from/to a file handle.
+    """
+
+    def __init__(self, message, line=None, column=None):
+        super(SerializerSyntaxError, self).__init__(message)
+        self.line = line
+        self.column = column
+
+    def __str__(self):
+        linestring = ' on line {}'.format(self.line) if self.line else ''
+        colstring = ' in column {}'.format(self.column) if self.column else ''
+        message = 'Syntax Error{}: {}'
+        return message.format(''.join([linestring, colstring]), self.message)

wlauto/result_processors/json_rp.py

@@ -0,0 +1,122 @@
+#    Copyright 2014-2015 ARM Limited
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+
+#pylint: disable=E1101,W0201
+import os
+from base64 import b64encode
+
+from wlauto import ResultProcessor, Parameter
+from wlauto.utils.serializer import json
+from wlauto.utils.misc import istextfile
+from wlauto.utils.types import list_of_strings
+from wlauto.exceptions import ResultProcessorError
+
+
+class JsonReportProcessor(ResultProcessor):
+
+    name = 'json'
+    description = """
+    Produces a JSON file with WA config, results ect.
+
+
+    This includes embedded artifacts either as text or base64
+
+    """
+
+    parameters = [
+        Parameter("ignore_artifact_types", kind=list_of_strings,
+                  default=['export', 'raw'],
+                  description="""A list of which artifact types to be ignored,
+                                 and thus not embedded in the JSON""")
+    ]
+    final = {}
+
+    def initialize(self, context):
+        self.final = context.run_info.to_dict()
+        del self.final['workload_specs']
+
+        wa_adapter = self.final['device']
+        self.final['device'] = {}
+        self.final['device']['props'] = self.final['device_properties']
+        self.final['device']['wa_adapter'] = wa_adapter
+        del self.final['device_properties']
+
+        self.final['output_directory'] = os.path.abspath(context.output_directory)
+        self.final['artifacts'] = []
+        self.final['workloads'] = context.config.to_dict()['workload_specs']
+        for workload in self.final['workloads']:
+            workload['name'] = workload['workload_name']
+            del workload['workload_name']
+            workload['results'] = []
+
+    def export_iteration_result(self, result, context):
+        r = {}
+        r['iteration'] = context.current_iteration
+        r['status'] = result.status
+        r['events'] = [e.to_dict() for e in result.events]
+        r['metrics'] = []
+        for m in result.metrics:
+            md = m.to_dict()
+            md['is_summary'] = m.name in context.workload.summary_metrics
+            r['metrics'].append(md)
+        iteration_artefacts = [self.embed_artifact(context, a) for a in context.iteration_artifacts]
+        r['artifacts'] = [e for e in iteration_artefacts if e is not None]
+        for workload in self.final['workloads']:
+            if workload['id'] == context.spec.id:
+                workload.update(r)
+                break
+        else:
+            raise ResultProcessorError("No workload spec with matching id found")
+
+    def export_run_result(self, result, context):
+        run_artifacts = [self.embed_artifact(context, a) for a in context.run_artifacts]
+        self.logger.debug('Generating results bundle...')
+        run_stats = {
+            'status': result.status,
+            'events': [e.to_dict() for e in result.events],
+            'end_time': context.run_info.end_time,
+            'duration': context.run_info.duration.total_seconds(),
+            'artifacts': [e for e in run_artifacts if e is not None],
+        }
+        self.final.update(run_stats)
+        json_path = os.path.join(os.path.abspath(context.output_directory), "run.json")
+        with open(json_path, 'w') as json_file:
+            json.dump(self.final, json_file)
+
+    def embed_artifact(self, context, artifact):
+        artifact_path = os.path.join(context.output_directory, artifact.path)
+
+        if not os.path.exists(artifact_path):
+            self.logger.debug('Artifact {} has not been generated'.format(artifact_path))
+            return
+        elif artifact.kind in self.ignore_artifact_types:
+            self.logger.debug('Ignoring {} artifact {}'.format(artifact.kind, artifact_path))
+            return
+        else:
+            self.logger.debug('Uploading artifact {}'.format(artifact_path))
+            entry = artifact.to_dict()
+            path = os.path.join(os.path.abspath(context.output_directory), entry['path'])
+            if istextfile(open(path)):
+                entry['encoding'] = "text"
+                entry['content'] = open(path).read()
+            else:
+                entry['encoding'] = "base64"
+                entry['content'] = b64encode(open(path).read())
+
+            del entry['path']
+            del entry['level']
+            del entry['mandatory']
+            return entry

wlauto/result_processors/standard.py

@@ -22,7 +22,6 @@ text files in various formats.
 """
 import os
 import csv
-import json
 
 from wlauto import ResultProcessor, Parameter
 from wlauto.exceptions import ConfigError
@@ -124,32 +123,6 @@ class CsvReportProcessor(ResultProcessor):
                     writer.writerow(row)
 
 
-class JsonReportProcessor(ResultProcessor):
-    """
-    Creates a ``results.json`` in the output directory containing results for
-    all iterations in JSON format.
-
-    """
-
-    name = 'json'
-
-    def process_run_result(self, result, context):
-        outfile = os.path.join(context.run_output_directory, 'results.json')
-        with open(outfile, 'wb') as wfh:
-            output = []
-            for result in result.iteration_results:
-                output.append({
-                    'id': result.id,
-                    'workload': result.workload.name,
-                    'iteration': result.iteration,
-                    'metrics': [dict([(k, v) for k, v in m.__dict__.iteritems()
-                                      if not k.startswith('_')])
-                                for m in result.metrics],
-                })
-            json.dump(output, wfh, indent=4)
-        context.add_artifact('run_result_json', 'results.json', 'export')
-
-
 class SummaryCsvProcessor(ResultProcessor):
     """
     Similar to csv result processor, but only contains workloads' summary metrics.

wlauto/utils/misc.py

@@ -815,3 +815,27 @@ def sha256(path, chunk=2048):
 
 def urljoin(*parts):
     return '/'.join(p.rstrip('/') for p in parts)
+
+
+# From: http://eli.thegreenplace.net/2011/10/19/perls-guess-if-file-is-text-or-binary-implemented-in-python/
+def istextfile(fileobj, blocksize=512):
+    """ Uses heuristics to guess whether the given file is text or binary,
+        by reading a single block of bytes from the file.
+        If more than 30% of the chars in the block are non-text, or there
+        are NUL ('\x00') bytes in the block, assume this is a binary file.
+    """
+    _text_characters = (b''.join(chr(i) for i in range(32, 127)) +
+                        b'\n\r\t\f\b')
+
+    block = fileobj.read(blocksize)
+    if b'\x00' in block:
+        # Files with null bytes are binary
+        return False
+    elif not block:
+        # An empty file is considered a valid text file
+        return True
+
+    # Use translate's 'deletechars' argument to efficiently remove all
+    # occurrences of _text_characters from the block
+    nontext = block.translate(None, _text_characters)
+    return float(len(nontext)) / len(block) <= 0.30

wlauto/utils/serializer.py

@@ -0,0 +1,243 @@
+"""
+This module contains wrappers for Python serialization modules for
+common formats that make it easier to serialize/deserialize WA
+Plain Old Data structures (serilizable WA classes implement
+``to_pod()``/``from_pod()`` methods for converting between POD
+structures and Python class instances).
+
+The modifications to standard serilization procedures are:
+
+    - mappings are deserialized as ``OrderedDict``\ 's are than standard
+      Python ``dict``\ 's. This allows for cleaner syntax in certain parts
+      of WA configuration (e.g. values to be written to files can be specified
+      as a dict, and they will be written in the order specified in the config).
+    - regular expressions are automatically encoded/decoded. This allows for
+      configuration values to be transparently specified as strings or regexes
+      in the POD config.
+
+This module exports the "wrapped" versions of serialization libraries,
+and this should be imported and used instead of importing the libraries
+directly. i.e. ::
+
+    from wa.utils.serializer import yaml
+    pod = yaml.load(fh)
+
+instead of ::
+
+    import yaml
+    pod = yaml.load(fh)
+
+It's also possible to use the serializer directly::
+
+    from wa.utils import serializer
+    pod = serializer.load(fh)
+
+This can also be used to ``dump()`` POD structures. By default,
+``dump()`` will produce JSON, but ``fmt`` parameter may be used to
+specify an alternative format (``yaml`` or ``python``). ``load()`` will
+use the file extension to guess the format, but ``fmt`` may also be used
+to specify it explicitly.
+
+"""
+# pylint: disable=unused-argument
+
+import os
+import re
+import json as _json
+from collections import OrderedDict
+from datetime import datetime
+
+import yaml as _yaml
+import dateutil.parser
+
+from wlauto.exceptions import SerializerSyntaxError
+from wlauto.utils.types import regex_type
+from wlauto.utils.misc import isiterable
+
+
+__all__ = [
+    'json',
+    'yaml',
+    'read_pod',
+    'dump',
+    'load',
+]
+
+
+class WAJSONEncoder(_json.JSONEncoder):
+
+    def default(self, obj):  # pylint: disable=method-hidden
+        if hasattr(obj, 'to_pod'):
+            return obj.to_pod()
+        elif isinstance(obj, regex_type):
+            return 'REGEX:{}:{}'.format(obj.flags, obj.pattern)
+        elif isinstance(obj, datetime):
+            return 'DATET:{}'.format(obj.isoformat())
+        else:
+            return _json.JSONEncoder.default(self, obj)
+
+
+class WAJSONDecoder(_json.JSONDecoder):
+
+    def decode(self, s, **kwargs):
+        d = _json.JSONDecoder.decode(self, s, **kwargs)
+
+        def try_parse_object(v):
+            if isinstance(v, basestring) and v.startswith('REGEX:'):
+                _, flags, pattern = v.split(':', 2)
+                return re.compile(pattern, int(flags or 0))
+            elif isinstance(v, basestring) and v.startswith('DATET:'):
+                _, pattern = v.split(':', 1)
+                return dateutil.parser.parse(pattern)
+            else:
+                return v
+
+        def load_objects(d):
+            pairs = []
+            for k, v in d.iteritems():
+                if hasattr(v, 'iteritems'):
+                    pairs.append((k, load_objects(v)))
+                elif isiterable(v):
+                    pairs.append((k, [try_parse_object(i) for i in v]))
+                else:
+                    pairs.append((k, try_parse_object(v)))
+            return OrderedDict(pairs)
+
+        return load_objects(d)
+
+
+class json(object):
+
+    @staticmethod
+    def dump(o, wfh, indent=4, *args, **kwargs):
+        return _json.dump(o, wfh, cls=WAJSONEncoder, indent=indent, *args, **kwargs)
+
+    @staticmethod
+    def load(fh, *args, **kwargs):
+        try:
+            return _json.load(fh, cls=WAJSONDecoder, object_pairs_hook=OrderedDict, *args, **kwargs)
+        except ValueError as e:
+            raise SerializerSyntaxError(e.message)
+
+    @staticmethod
+    def loads(s, *args, **kwargs):
+        try:
+            return _json.loads(s, cls=WAJSONDecoder, object_pairs_hook=OrderedDict, *args, **kwargs)
+        except ValueError as e:
+            raise SerializerSyntaxError(e.message)
+
+
+_mapping_tag = _yaml.resolver.BaseResolver.DEFAULT_MAPPING_TAG
+_regex_tag = u'tag:wa:regex'
+
+
+def _wa_dict_representer(dumper, data):
+    return dumper.represent_mapping(_mapping_tag, data.iteritems())
+
+
+def _wa_regex_representer(dumper, data):
+    text = '{}:{}'.format(data.flags, data.pattern)
+    return dumper.represent_scalar(_regex_tag, text)
+
+
+def _wa_dict_constructor(loader, node):
+    pairs = loader.construct_pairs(node)
+    seen_keys = set()
+    for k, _ in pairs:
+        if k in seen_keys:
+            raise ValueError('Duplicate entry: {}'.format(k))
+        seen_keys.add(k)
+    return OrderedDict(pairs)
+
+
+def _wa_regex_constructor(loader, node):
+    value = loader.construct_scalar(node)
+    flags, pattern = value.split(':', 1)
+    return re.compile(pattern, int(flags or 0))
+
+
+_yaml.add_representer(OrderedDict, _wa_dict_representer)
+_yaml.add_representer(regex_type, _wa_regex_representer)
+_yaml.add_constructor(_mapping_tag, _wa_dict_constructor)
+_yaml.add_constructor(_regex_tag, _wa_regex_constructor)
+
+
+class yaml(object):
+
+    @staticmethod
+    def dump(o, wfh, *args, **kwargs):
+        return _yaml.dump(o, wfh, *args, **kwargs)
+
+    @staticmethod
+    def load(fh, *args, **kwargs):
+        try:
+            return _yaml.load(fh, *args, **kwargs)
+        except _yaml.YAMLError as e:
+            lineno = None
+            if hasattr(e, 'problem_mark'):
+                lineno = e.problem_mark.line  # pylint: disable=no-member
+            raise SerializerSyntaxError(e.message, lineno)
+
+    loads = load
+
+
+class python(object):
+
+    @staticmethod
+    def dump(o, wfh, *args, **kwargs):
+        raise NotImplementedError()
+
+    @classmethod
+    def load(cls, fh, *args, **kwargs):
+        return cls.loads(fh.read())
+
+    @staticmethod
+    def loads(s, *args, **kwargs):
+        pod = {}
+        try:
+            exec s in pod  # pylint: disable=exec-used
+        except SyntaxError as e:
+            raise SerializerSyntaxError(e.message, e.lineno)
+        for k in pod.keys():
+            if k.startswith('__'):
+                del pod[k]
+        return pod
+
+
+def read_pod(source, fmt=None):
+    if isinstance(source, basestring):
+        with open(source) as fh:
+            return _read_pod(fh, fmt)
+    elif hasattr(source, 'read') and (hasattr(source, 'name') or fmt):
+        return _read_pod(source, fmt)
+    else:
+        message = 'source must be a path or an open file handle; got {}'
+        raise ValueError(message.format(type(source)))
+
+
+def dump(o, wfh, fmt='json', *args, **kwargs):
+    serializer = {'yaml': yaml,
+                  'json': json,
+                  'python': python,
+                  'py': python,
+                  }.get(fmt)
+    if serializer is None:
+        raise ValueError('Unknown serialization format: "{}"'.format(fmt))
+    serializer.dump(o, wfh, *args, **kwargs)
+
+
+def load(s, fmt='json', *args, **kwargs):
+    return read_pod(s, fmt=fmt)
+
+
+def _read_pod(fh, fmt=None):
+    if fmt is None:
+        fmt = os.path.splitext(fh.name)[1].lower().strip('.')
+    if fmt == 'yaml':
+        return yaml.load(fh)
+    elif fmt == 'json':
+        return json.load(fh)
+    elif fmt == 'py':
+        return python.load(fh)
+    else:
+        raise ValueError('Unknown format "{}": {}'.format(fmt, getattr(fh, 'name', '<none>')))