Merge pull request #353 from delph-in/release-1.7.0

Release 1.7.0

.github/workflows/checks.yml

@@ -12,7 +12,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: [3.6, 3.7, 3.8, 3.9]
+        python-version: ["3.7", "3.8", "3.9", "3.10"]
 
     steps:
     - uses: actions/checkout@v2

CHANGELOG.md

@@ -1,5 +1,35 @@
 # Change Log
 
+## [v1.7.0]
+
+**Release date: 2022-10-13**
+
+Updates Python support to currently active releases. Includes
+`delphin.highlight` (adds a dependency on Pygments), `delphin.edm`,
+and `delphin.codecs.dmrstikz` directly instead of as separate plugins.
+
+### Python Versions
+
+* Removed Python 3.6 support ([#343])
+* Added Python 3.10 support ([#343])
+
+### Added
+
+* `delphin.highlight` for syntax highlighting; it is no longer
+  necessary to install this as a plugin ([#344])
+* `delphin.edm` for elementary dependency matching; it is no longer
+  necessary to install this as a plugin ([#344])
+* `delphin.codecs.dmrstikz` for exporting DMRS with tikz-dependency;
+  it is no longer necessary to install this as a plugin ([#344])
+* `executable` parameter to `delphin.commands.process()` ([#352])
+* `--executable` option to `delphin process` command ([#352])
+
+### Fixed
+
+* The `mkprof` command no longer raises an error on an empty sentence ([#335])
+* EDS conversion in `delphin.edm` now more robustly handles errors
+
+
 ## [v1.6.0]
 
 **Release date: 2021-09-30**
@@ -1346,6 +1376,7 @@ There was no CHANGELOG file prior to this release, so I don't have much
 information about changes, except for
 [commit messages](../../commits/v0.2).
 
+[v1.7.0]: ../../releases/tag/v1.7.0
 [v1.6.0]: ../../releases/tag/v1.6.0
 [v1.5.1]: ../../releases/tag/v1.5.1
 [v1.5.0]: ../../releases/tag/v1.5.0
@@ -1521,3 +1552,7 @@ information about changes, except for
 [#331]: https://github.com/delph-in/pydelphin/issues/331
 [#333]: https://github.com/delph-in/pydelphin/issues/333
 [#334]: https://github.com/delph-in/pydelphin/issues/334
+[#335]: https://github.com/delph-in/pydelphin/issues/335
+[#343]: https://github.com/delph-in/pydelphin/issues/343
+[#344]: https://github.com/delph-in/pydelphin/issues/344
+[#352]: https://github.com/delph-in/pydelphin/issues/352

CONTRIBUTING.md

@@ -27,12 +27,12 @@ For bug requests, please provide the following, if possible:
   ```python
   >>> from delphin.__about__ import __version__
   >>> __version__  # distribution version
-  '1.6.0'
+  '1.7.0'
   >>> from delphin import mrs
   >>> mrs.__version__  # package version
-  '1.6.0'
+  '1.7.0'
   ```
-* Python version (e.g. 3.6, 3.7, etc.)
+* Python version (e.g. 3.7, 3.8, etc.)
 
 For feature requests, please provide a use case for the feature.
 

delphin/__about__.py

@@ -3,7 +3,7 @@
 # the warehouse project:
 #   https://github.com/pypa/warehouse/blob/master/warehouse/__about__.py
 
-__version__ = '1.6.0'
+__version__ = '1.7.0'
 __version_info__ = __version__.replace('.', ' ').replace('-', ' ').split()
 
 __title__ = 'PyDelphin'

delphin/cli/convert.py

@@ -32,17 +32,6 @@ def call_convert(args):
     else:
         color = (args.color == 'always'
                  or (args.color == 'auto' and sys.stdout.isatty()))
-        if color:
-            try:
-                import delphin.highlight  # noqa: F401
-            except ImportError:
-                # don't warn if color=auto
-                if args.color == 'always':
-                    warnings.warn(
-                        'delphin.highlight must be installed for '
-                        'syntax highlighting',
-                        PyDelphinWarning)
-                color = False
         if args.indent and args.indent is not True:
             if args.indent.lower() in ('no', 'none'):
                 args.indent = None

delphin/cli/edm.py

@@ -0,0 +1,152 @@
+
+"""
+Compute the EDM (Elementary Dependency Match) score for two collections.
+
+The collections, GOLD and TEST, may be files containing semantic
+representations, decoded using '--format'; or an [incr tsdb()] test
+suite directory, selecting MRSs using '-p' for the parse result
+number. GOLD and TEST should contain the same number of items. MRS
+representations will be converted to EDS for comparison.
+"""
+
+from typing import Optional, Iterator, Union
+import argparse
+from pathlib import Path
+import warnings
+import logging
+
+from delphin import util
+from delphin import tsdb
+from delphin import itsdb
+from delphin import edm
+from delphin import mrs
+from delphin import eds
+from delphin import dmrs
+
+
+logger = logging.getLogger(__name__)
+
+_SemanticRepresentation = Union[eds.EDS, dmrs.DMRS]
+
+parser = argparse.ArgumentParser(add_help=False)
+
+COMMAND_INFO = {
+    'name': 'edm',
+    'help': 'Evaluate with Elementary Dependency Match',
+    'description': __doc__,
+    'parser': parser,
+}
+
+
+def call_compute(args):
+    golds = _iter_representations(args.GOLD, args.format, args.p)
+    tests = _iter_representations(args.TEST, args.format, args.p)
+    p, r, f = edm.compute(
+        golds,
+        tests,
+        name_weight=args.N,
+        argument_weight=args.A,
+        property_weight=args.P,
+        constant_weight=args.C,
+        top_weight=args.T,
+        ignore_missing_gold=args.ignore_missing in ('gold', 'both'),
+        ignore_missing_test=args.ignore_missing in ('test', 'both'))
+    print(f'Precision:\t{p}')
+    print(f'   Recall:\t{r}')
+    print(f'  F-score:\t{f}')
+
+
+def _iter_representations(
+    path: Path,
+    fmt: str,
+    p: int
+) -> Iterator[Optional[_SemanticRepresentation]]:
+    if tsdb.is_database_directory(path):
+        logger.debug('reading MRSs from profile: %s', (path,))
+        ts = itsdb.TestSuite(path)
+        for response in ts.processed_items():
+            try:
+                result = response.result(p)
+            except IndexError:
+                yield None
+            else:
+                yield _eds_from_mrs(result.mrs(), predicate_modifiers=True)
+
+    elif path.is_file():
+        logger.debug('reading %s from file: %s', (fmt, path,))
+        codec = util.import_codec(fmt)
+        rep = codec.CODEC_INFO.get('representation', '').lower()
+        if rep == 'mrs':
+            for sr in codec.load(path):
+                yield _eds_from_mrs(sr, predicate_modifiers=True)
+        elif rep in ('dmrs', 'eds'):
+            for sr in codec.load(path):
+                yield sr
+        else:
+            raise ValueError(f'unsupported representation: {rep}')
+
+    else:
+        raise ValueError(f'not a file or TSDB database: {path}')
+
+
+def _eds_from_mrs(
+    m: mrs.MRS,
+    predicate_modifiers: bool,
+    errors: str = 'warn',
+) -> Optional[eds.EDS]:
+    try:
+        e = eds.from_mrs(m, predicate_modifiers=predicate_modifiers)
+    except Exception:
+        logger.debug('could not convert MRS to EDS')
+        if errors == 'warn':
+            warnings.warn("error in EDS conversion; skipping entry")
+        elif errors == 'strict':
+            raise
+        e = None
+    return e
+
+
+parser.set_defaults(func=call_compute)
+# data selection
+parser.add_argument(
+    'GOLD',
+    type=Path,
+    help='corpus of gold semantic representations')
+parser.add_argument(
+    'TEST',
+    type=Path,
+    help='corpus of test semantic representations')
+parser.add_argument(
+    '-f',
+    '--format',
+    metavar='FMT',
+    default='eds',
+    help='semantic representation format (default: eds)')
+parser.add_argument(
+    '-p',
+    metavar='N',
+    type=int,
+    default=0,
+    help='parse result number (default: 0)')
+parser.add_argument(
+    '--ignore-missing',
+    metavar='X',
+    choices=('gold', 'test', 'both', 'none'),
+    default='none',
+    help='do not treat missing Xs as a mismatch (default: none)')
+# comparison configuration
+parser.add_argument(
+    '-A', metavar='WEIGHT', type=float, default=1.0,
+    help='weight for argument triples (default: 1.0)')
+parser.add_argument(
+    '-N', metavar='WEIGHT', type=float, default=1.0,
+    help='weight for name (predicate) triples (default: 1.0)')
+parser.add_argument(
+    '-P', metavar='WEIGHT', type=float, default=1.0,
+    help='weight for property triples (default: 1.0)')
+parser.add_argument(
+    '-C', metavar='WEIGHT', type=float, default=1.0,
+    help='weight for constant triples (default: 1.0)')
+parser.add_argument(
+    '-T', metavar='WEIGHT', type=float, default=1.0,
+    help='weight for matching top triples (default: 1.0)')

delphin/cli/process.py

@@ -43,7 +43,8 @@ def call_process(args):
         options=shlex.split(args.options),
         all_items=args.all_items,
         result_id=args.p,
-        gzip=args.gzip)
+        gzip=args.gzip,
+        executable=args.executable)
 
 
 # process subparser
@@ -92,3 +93,6 @@ def call_process(args):
 )
 parser.add_argument(
     '-z', '--gzip', action='store_true', help='compress table files with gzip')
+parser.add_argument(
+    '--executable', metavar='PATH', default='ace',
+    help='path to ACE executable (default: ace)')

delphin/cli/repp.py

@@ -30,16 +30,6 @@
 def call_repp(args):
     color = (args.color == 'always'
              or (args.color == 'auto' and sys.stdout.isatty()))
-    if color:
-        try:
-            import pygments  # noqa: F401
-        except ImportError:
-            # don't warn if color=auto
-            if args.color == 'always':
-                warnings.warn(
-                    'Pygments must be installed for diff highlighting',
-                    PyDelphinWarning)
-            color = False
     return repp(
         args.FILE or sys.stdin,
         config=args.config,

delphin/codecs/ace.py

@@ -1,5 +1,3 @@
-# -*- coding: utf-8 -*-
-
 """
 Deserialization of MRSs in ACE's stdout protocols.
 """

delphin/codecs/dmrsjson.py

@@ -1,5 +1,3 @@
-# -*- coding: utf-8 -*-
-
 """
 DMRS-JSON serialization and deserialization.
 """

delphin/codecs/dmrspenman.py

@@ -1,5 +1,3 @@
-# -*- coding: utf-8 -*-
-
 """
 DMRS-PENMAN serialization and deserialization.
 """