Add more specialized exceptions

Resolves #198

CHANGELOG.md

@@ -27,6 +27,7 @@ these changes are prefixed with "**BREAKING**"
   - `STRING_TYPE`
   - `SemI.find_synopsis()`
   - `SemIError`
+  - `SemISyntaxError`
   - `SemIWarning`
   - `Synopsis`
   - `SynopsisRole`
@@ -45,6 +46,7 @@ these changes are prefixed with "**BREAKING**"
   - `EVENTUALITY`
   - `INSTANCE`
   - `is_valid()`
+* `delphin.mrs.vpm.VPMSyntaxError`
 
 ### Moved or Renamed
 
@@ -55,6 +57,7 @@ these changes are prefixed with "**BREAKING**"
 * `delphin.exceptions.TdlWarning` to `delphin.tdl.TDLWarning`
 * `delphin.exceptions.TSQLSyntaxError` to `delphin.tsql.TSQLSyntaxError`
 * `delphin.extra.highlight.TdlLexer` to `delphin.extra.highlight.TDLLexer`
+* `delphin.interfaces.ace.AceProcessError` to `delphin.interfaces.ace.ACEProcessError`
 * `delphin.mrs.util.rargname_sortkey()` to `delphin.sembase.role_priority()`
 * `delphin.mrs.components.Lnk` to `delphin.lnk.Lnk`
 * `delphin.mrs.components._LnkMixin` to `delphin.lnk.LnkMixin`
@@ -115,12 +118,17 @@ these changes are prefixed with "**BREAKING**"
 ### Changed
 
 * `delphin.extra.highlight` add docstrings and wild-card to TDL highlighter
+* `delphin.interfaces.ace` no longer raise `ValueError` when grammar file does
+  not exist; the `ACEProcessError` now handles this
 * `delphin.semi.load()` takes an `encoding` parameter
 * `delphin.semi` dictionary schema removes empty/default values; changes
   structure of predicate synopses
+* `delphin.semi` warns when a type is redefined
 * `delphin.tfs` -- errors raise `TypeHierarchyError` instead of `ValueError`
 * `delphin.tfs` nodes in the hierarchy are now TypeHierarchyNodes and may
   contain arbitrary data in addition to the parents and children
+* `delphin.mrs.vpm` raises `VPMSyntaxError` on bad inputs instead of
+  `AssertionError`
 
 
 ## [v0.9.2][]

delphin/interfaces/ace.py

@@ -83,7 +83,7 @@
 from delphin.exceptions import PyDelphinException
 
 
-class AceProcessError(PyDelphinException):
+class ACEProcessError(PyDelphinException):
     """Raised when the ACE process has crashed and cannot be recovered."""
 
 
@@ -120,8 +120,6 @@ class AceProcess(Processor):
 
     def __init__(self, grm, cmdargs=None, executable=None, env=None,
                  tsdbinfo=True, **kwargs):
-        if not os.path.isfile(grm):
-            raise ValueError("Grammar file %s does not exist." % grm)
         self.grm = grm
 
         self.cmdargs = cmdargs or []
@@ -172,7 +170,7 @@ def _open(self):
             'start': datetime.now()
         })
         if self._p.poll() is not None and self._p.returncode != 0:
-            raise AceProcessError('Process closed on startup; see <stderr>.')
+            raise ACEProcessError('Process closed on startup; see <stderr>.')
 
     def __enter__(self):
         return self

delphin/mrs/vpm.py

@@ -14,12 +14,18 @@
 
 import re
 
+from delphin.exceptions import PyDelphinSyntaxError
 from delphin import variable
 
 _LR_OPS = set(['<>', '>>', '==', '=>'])
 _RL_OPS = set(['<>', '<<', '==', '<='])
 _SUBSUME_OPS = set(['<>', '<<', '>>'])
 _EQUAL_OPS = set(['==', '<=', '=>'])
+_ALL_OPS = _LR_OPS.union(_RL_OPS)
+
+
+class VPMSyntaxError(PyDelphinSyntaxError):
+    """Raised when loading an invalid VPM."""
 
 
 def load(source, semi=None):
@@ -41,11 +47,12 @@ def load(source, semi=None):
             return _load(fh, semi)
 
 def _load(fh, semi):
+    filename = getattr(fh, 'name', '<stream>')
     typemap = []
     propmap = []
     curmap = typemap
     lh, rh = 1, 1  # number of elements expected on each side
-    for line in fh:
+    for lineno, line in enumerate(fh, 1):
         line = line.lstrip()
 
         if not line or line.startswith(';'):
@@ -63,16 +70,31 @@ def _load(fh, semi):
         match = re.match(r'(?P<lvals>.*)(?P<op>[<>=]{2})(?P<rvals>.*)$', line)
         if match is not None:
             lvals = match.group('lvals').split()
+            op = match.group('op')
             rvals = match.group('rvals').split()
-            assert len(lvals) == lh
-            assert len(rvals) == rh
-            curmap.append((lvals, match.group('op'), rvals))
+            msg, offset = '', -1
+            if len(lvals) != lh:
+                msg = 'wrong number of values on left side'
+                offset = match.end('lvals')
+            if op not in _ALL_OPS:
+                msg = 'invalid operator'
+                offset = match.end('op')
+            elif len(rvals) != rh:
+                msg = 'wrong number of values on right side'
+                offset = match.end('rvals')
+            if msg:
+                raise VPMSyntaxError(msg, filename=filename,
+                                     lineno=lineno, offset=offset,
+                                     text=line)
+            curmap.append((lvals, op, rvals))
             continue
 
-        raise ValueError('Invalid line in VPM file: {}'.format(line))
+        raise VPMSyntaxError('invalid line in VPM file',
+                             filename=filename, lineno=lineno, text=line)
 
     return VPM(typemap, propmap, semi)
 
+
 class VPM(object):
     """
     A variable-property mapping.

delphin/semi.py

@@ -24,6 +24,7 @@
 from delphin import tfs
 from delphin.exceptions import (
     PyDelphinException,
+    PyDelphinSyntaxError,
     PyDelphinWarning
 )
 
@@ -78,15 +79,15 @@
 
 
 class SemIError(PyDelphinException):
-    """
-    Raised when loading an invalid SEM-I.
-    """
+    """Raised when loading an invalid SEM-I."""
+
+
+class SemISyntaxError(PyDelphinSyntaxError):
+    """Raised when loading an invalid SEM-I."""
 
 
 class SemIWarning(PyDelphinWarning):
-    """
-    Warning class for questionable SEM-Is.
-    """
+    """Warning class for questionable SEM-Is."""
 
 
 def load(fn, encoding='utf-8'):
@@ -107,14 +108,14 @@ def load(fn, encoding='utf-8'):
 
 def _read_file(fn, basedir, encoding):
     data = {
-        'variables': [],
-        'properties': [],
-        'roles': [],
+        'variables': {},
+        'properties': {},
+        'roles': {},
         'predicates': {},
     }
     section = None
 
-    for line in open(fn, 'r', encoding=encoding):
+    for lineno, line in enumerate(open(fn, 'r', encoding=encoding), 1):
         line = line.lstrip()
 
         if not line or line.startswith(';'):
@@ -124,7 +125,9 @@ def _read_file(fn, basedir, encoding):
         if match is not None:
             name = match.group('name')
             if name not in _SEMI_SECTIONS:
-                raise ValueError('Invalid SEM-I section: {}'.format(name))
+                raise SemISyntaxError(
+                    'invalid SEM-I section',
+                    filename=fn, lineno=lineno, text=line)
             else:
                 section = name
             continue
@@ -134,9 +137,12 @@ def _read_file(fn, basedir, encoding):
             include_fn = pjoin(basedir, match.group('filename').rstrip())
             include_data = _read_file(
                 include_fn, dirname(include_fn), encoding)
-            data['variables'].extend(include_data.get('variables', []))
-            data['properties'].extend(include_data.get('properties', []))
-            data['roles'].extend(include_data.get('roles', []))
+            for key, val in include_data['variables'].items():
+                _incorporate(data['variables'], key, val, include_fn)
+            for key, val in include_data['properties'].items():
+                _incorporate(data['properties'], key, val, include_fn)
+            for key, val in include_data['roles'].items():
+                _incorporate(data['roles'], key, val, include_fn)
             for pred, d in include_data['predicates'].items():
                 if pred not in data['predicates']:
                     data['predicates'][pred] = {
@@ -148,7 +154,7 @@ def _read_file(fn, basedir, encoding):
                 if d.get('synopses'):
                     data['predicates'][pred]['synopses'].extend(d['synopses'])
 
-        if section == 'variables':
+        elif section == 'variables':
             # e.g. e < i : PERF bool, TENSE tense.
             match = _variable_entry_re.match(line)
             if match is not None:
@@ -162,9 +168,11 @@ def _read_file(fn, basedir, encoding):
                     properties = [pair.split() for pair in pairs]
                 v = {'parents': supertypes, 'properties': properties}
                 # v = type(identifier, supertypes, d)
-                data['variables'].append((identifier, v))
+                _incorporate(data['variables'], identifier, v, fn)
             else:
-                raise ValueError('Invalid entry: {}'.format(line))
+                raise SemISyntaxError(
+                    'invalid variable',
+                    filename=fn, lineno=lineno, text=line)
 
         elif section == 'properties':
             # e.g. + < bool.
@@ -174,18 +182,23 @@ def _read_file(fn, basedir, encoding):
                 supertypes = match.group('parents') or []
                 if supertypes:
                     supertypes = supertypes.split(' & ')
-                data['properties'].append((_type, {'parents': supertypes}))
+                _incorporate(
+                    data['properties'], _type, {'parents': supertypes}, fn)
             else:
-                raise ValueError('Invalid entry: {}'.format(line))
+                raise SemISyntaxError(
+                    'invalid property',
+                    filename=fn, lineno=lineno, text=line)
 
         elif section == 'roles':
             # e.g. + < bool.
             match = _role_entry_re.match(line)
             if match is not None:
-                rargname, value = match.group('role'), match.group('value')
-                data['roles'].append((rargname, {'value': value}))
+                role, value = match.group('role'), match.group('value')
+                _incorporate(data['roles'], role, {'value': value}, fn)
             else:
-                raise ValueError('Invalid entry: {}'.format(line))
+                raise SemISyntaxError(
+                    'invalid role',
+                    filename=fn, lineno=lineno, text=line)
 
         elif section == 'predicates':
             # e.g. _predicate_n_1 : ARG0 x { IND + }.
@@ -217,6 +230,12 @@ def _read_file(fn, basedir, encoding):
     return data
 
 
+def _incorporate(d, key, val, fn):
+    if key in d:
+        warnings.warn("'{}' redefined in {}".format(key, fn), SemIWarning)
+    d[key] = val
+
+
 class SynopsisRole(tuple):
     """
     Role data associated with a SEM-I predicate synopsis.
@@ -358,22 +377,26 @@ def __init__(self,
         self.roles = {}
         self.predicates = tfs.TypeHierarchy(TOP_TYPE)
         # validate and normalize inputs
-        self._init_properties(properties)
-        self._init_variables(variables)
-        self._init_roles(roles)
-        self._init_predicates(predicates)
+        if properties:
+            self._init_properties(properties)
+        if variables:
+            self._init_variables(variables)
+        if roles:
+            self._init_roles(roles)
+        if predicates:
+            self._init_predicates(predicates)
 
     def _init_properties(self, properties):
         subhier = {}
-        for prop, data in dict(properties or []).items():
+        for prop, data in properties.items():
             prop = prop.lower()
             parents = data.get('parents')
             _add_to_subhierarchy(subhier, prop, parents, None)
         self.properties.update(subhier)
 
     def _init_variables(self, variables):
         subhier = {}
-        for var, data in dict(variables or []).items():
+        for var, data in variables.items():
             var = var.lower()
             parents = data.get('parents')
             properties = []
@@ -386,7 +409,7 @@ def _init_variables(self, variables):
         self.variables.update(subhier)
 
     def _init_roles(self, roles):
-        for role, data in dict(roles or []).items():
+        for role, data in roles.items():
             role = role.upper()
             var = data['value'].lower()
             if not (var == STRING_TYPE or var in self.variables):
@@ -396,7 +419,7 @@ def _init_roles(self, roles):
     def _init_predicates(self, predicates):
         subhier = {}
         propcache = {v: dict(node.data) for v, node in self.variables.items()}
-        for pred, data in dict(predicates or []).items():
+        for pred, data in predicates.items():
             pred = pred.lower()
             parents = data.get('parents')
             synopses = []

docs/api/delphin.interfaces.ace.rst

@@ -56,5 +56,5 @@ delphin.interfaces.ace
   Exceptions
   ----------
 
-  .. autoexception:: AceProcessError
+  .. autoexception:: ACEProcessError
      :show-inheritance:

docs/api/delphin.mrs.vpm.rst

@@ -4,6 +4,19 @@ delphin.mrs.vpm
 
 .. automodule:: delphin.mrs.vpm
 
-  .. autofunction:: load
-  .. autoclass:: VPM
-    :members:
+   Module functions
+   ----------------
+
+   .. autofunction:: load
+
+   Classes
+   -------
+
+   .. autoclass:: VPM
+      :members:
+
+   Exceptions
+   ----------
+
+   .. autoexception:: VPMSyntaxError
+      :show-inheritance:

docs/api/delphin.semi.rst

@@ -61,8 +61,14 @@ delphin.semi
 
   .. autoclass:: SynopsisRole
 
-  Exceptions
-  ----------
+  Exceptions and Warnings
+  -----------------------
 
   .. autoexception:: SemIError
     :show-inheritance:
+
+  .. autoexception:: SemISyntaxError
+    :show-inheritance:
+
+  .. autoexception:: SemIWarning
+    :show-inheritance:

tests/interfaces_ace_test.py

@@ -53,7 +53,7 @@ def test_start(ace_mismatch, tmpdir, monkeypatch):
     with monkeypatch.context() as m:
         m.setattr(ace, 'Popen', popen)
         m.setattr(ace, '_ace_version', lambda x: (0, 9, 29))
-        with pytest.raises(ace.AceProcessError):
+        with pytest.raises(ace.ACEProcessError):
             ace.AceParser(str(grm))
-        with pytest.raises(ace.AceProcessError):
+        with pytest.raises(ace.ACEProcessError):
             ace.parse(str(grm), 'Dogs sleep.')