diff --git a/app/corpora/cqi/__init__.py b/app/corpora/cqi/__init__.py
index e543ac6ee5d9c86773fac03a9de01d89759f4520..4174f55dbadaab06c0e80d93433cb82042b54e29 100644
--- a/app/corpora/cqi/__init__.py
+++ b/app/corpora/cqi/__init__.py
@@ -1,7 +1,6 @@
# flake8: noqa
from .api import APIClient
from .client import CQiClient
-from .wrapper import CQiWrapper
from .version import version, version_info
diff --git a/app/corpora/cqi/api/client.py b/app/corpora/cqi/api/client.py
index 9c25720378d66c4a5955dec85c85cecfaa9125d2..bc62e65f786f8a4a4a943670bfd0e4f02075323f 100644
--- a/app/corpora/cqi/api/client.py
+++ b/app/corpora/cqi/api/client.py
@@ -20,10 +20,11 @@ class APIClient:
>>> client.ctrl_bye()
{'code': 259, 'msg': 'CQI_STATUS_BYE_OK'}
- Args:
+ Attributes:
host (str): URL to the CQP server. For example,
``cqpserver.localhost`` or ``127.0.0.1``.
port (int): Port the CQP server listens on. Default: ``4877``
+ socket (socket.socket): Socket for communicating with a CQP server.
"""
def __init__(self, host, port=4877):
diff --git a/app/corpora/cqi/client.py b/app/corpora/cqi/client.py
index 5cc867c1a63f4b7cfde4a77be7c84ce25fa6ddd4..0fba29eb71680f2180149f8f039fa539fc700e69 100644
--- a/app/corpora/cqi/client.py
+++ b/app/corpora/cqi/client.py
@@ -15,8 +15,9 @@ class CQiClient:
{'code': 260, 'msg': 'CQI_STATUS_PING_OK'}
>>> client.disconnect()
{'code': 259, 'msg': 'CQI_STATUS_BYE_OK'}
+
Attributes:
- api (APIClient): A client pointing to the specified to the CQP server.
+ api (APIClient): An API client pointing to the specified CQP server.
"""
def __init__(self, host, port=4877):
@@ -32,12 +33,14 @@ class CQiClient:
def connect(self, username='anonymous', password=''):
status = self.api.ctrl_connect(username, password)
- self.corpora = CorpusCollection(self)
return status
def disconnect(self):
- del self.corpora
return self.api.ctrl_bye()
def ping(self):
return self.api.ctrl_ping()
+
+ @property
+ def corpora(self):
+ return CorpusCollection(client=self)
diff --git a/app/corpora/cqi/models/attributes.py b/app/corpora/cqi/models/attributes.py
index 20dd359b43a25ee1723e11510d9c0960c32e596c..ea92d5af8caddee90ca3cc0c37468b4ce7ac36e6 100644
--- a/app/corpora/cqi/models/attributes.py
+++ b/app/corpora/cqi/models/attributes.py
@@ -1,139 +1,119 @@
-class AttributeCollection:
- def __init__(self, client, corpus):
- self.client = client
- self.corpus = corpus
- self.alignment = AlignmentAttributeCollection(client, corpus)
- self.positional = PositionalAttributeCollection(client, corpus)
- self.structural = StructuralAttributeCollection(client, corpus)
+from .ressource import Collection, Model
-class Attribute:
+class Attribute(Model):
"""
This is a class representing an attribute. Attributes denote the general
category of information. A specific occurence is identified by an Id.
-
- Attributes:
- client (CQiClient): A connected client pointing at the server that this
- object is on.
- corpus (Corpus): The corpus, this attribute belongs to.
- name (str): The name of the Attribute.
- size (int): The number of occurences of this attribute within the corpus.
"""
- def __init__(self, client, corpus, name):
- self.client = client
- self.corpus = corpus
- self.name = name
- self._name = '{}.{}'.format(corpus.name, name)
- self.size = client.api.cl_attribute_size(self._name)
+ id_attribute = 'api_name'
+
+ @staticmethod
+ def _attrs(client, corpus, name):
+ api_name = '{}.{}'.format(corpus.attrs['api_name'], name)
+ return {'api_name': api_name,
+ 'name': name,
+ 'size': client.api.cl_attribute_size(api_name)}
def drop(self):
- return self.client.api.cl_drop_attribute(self._name)
+ return self.client.api.cl_drop_attribute(self.attrs['api_name'])
+
+class AttributeCollection(Collection):
+ model = Attribute
-class AlignmentAttributeCollection:
- def __init__(self, client, corpus):
- self.client = client
+ def __init__(self, client=None, corpus=None):
+ super(AttributeCollection, self).__init__(client=client)
self.corpus = corpus
- def get(self, name):
- return AlignmentAttribute(self.client, self.corpus, name)
+ def get(self, attribute_name):
+ return self.prepare_model(self.model._attrs(self.client, self.corpus,
+ attribute_name))
def list(self):
- return [AlignmentAttribute(self.client, self.corpus, attr) for attr in
- self.client.api.corpus_alignment_attributes(self.corpus.name)]
+ raise NotImplementedError
class AlignmentAttribute(Attribute):
def cpos_by_ids(self, id_list):
- return self.client.api.cl_alg2cpos(self._name, id_list)
+ return self.client.api.cl_alg2cpos(self.attrs['api_name'], id_list)
def ids_by_cpos(self, cpos_list):
- return self.client.api.cl_cpos2alg(self._name, cpos_list)
-
+ return self.client.api.cl_cpos2alg(self.attrs['api_name'], cpos_list)
-class PositionalAttributeCollection:
- def __init__(self, client, corpus):
- self.client = client
- self.corpus = corpus
- def get(self, name):
- return PositionalAttribute(self.client, self.corpus, name)
+class AlignmentAttributeCollection(AttributeCollection):
+ model = AlignmentAttribute
def list(self):
- return [PositionalAttribute(self.client, self.corpus, attr) for attr in
- self.client.api.corpus_positional_attributes(self.corpus.name)]
+ return [self.prepare_model(self.model._attrs(self.client, self.corpus, attr)) # noqa
+ for attr in self.client.api.corpus_alignment_attributes(self.corpus.attrs['api_name'])] # noqa
class PositionalAttribute(Attribute):
- def __init__(self, client, corpus, name):
- super(PositionalAttribute, self).__init__(client, corpus, name)
- self.lexicon_size = client.api.cl_lexicon_size(self._name)
+ @staticmethod
+ def _attrs(client, corpus, name):
+ attrs = super(PositionalAttribute, PositionalAttribute)._attrs(client, corpus, name) # noqa
+ attrs['lexicon_size'] = client.api.cl_lexicon_size(attrs['api_name'])
+ return attrs
def cpos_by_id(self, id):
- return self.client.api.cl_id2cpos(self._name, id)
+ return self.client.api.cl_id2cpos(self.attrs['api_name'], id)
def cpos_by_ids(self, id_list):
- return self.client.api.cl_idlist2cpos(self._name, id_list)
+ return self.client.api.cl_idlist2cpos(self.attrs['api_name'], id_list)
def freqs_by_ids(self, id_list):
- return self.client.api.cl_id2freq(self._name, id_list)
+ return self.client.api.cl_id2freq(self.attrs['api_name'], id_list)
def ids_by_cpos(self, cpos_list):
- return self.client.api.cl_cpos2id(self._name, cpos_list)
+ return self.client.api.cl_cpos2id(self.attrs['api_name'], cpos_list)
def ids_by_regex(self, regex):
- return self.client.api.cl_regex2id(self._name, regex)
+ return self.client.api.cl_regex2id(self.attrs['api_name'], regex)
def ids_by_values(self, value_list):
- return self.client.api.cl_str2id(self._name, value_list)
+ return self.client.api.cl_str2id(self.attrs['api_name'], value_list)
def values_by_cpos(self, cpos_list):
- return self.client.api.cl_cpos2str(self._name, cpos_list)
+ return self.client.api.cl_cpos2str(self.attrs['api_name'], cpos_list)
def values_by_ids(self, id_list):
- return self.client.api.cl_id2str(self._name, id_list)
-
+ return self.client.api.cl_id2str(self.attrs['api_name'], id_list)
-class StructuralAttributeCollection:
- def __init__(self, client, corpus):
- self.client = client
- self.corpus = corpus
- def get(self, name):
- return StructuralAttribute(self.client, self.corpus, name)
+class PositionalAttributeCollection(AttributeCollection):
+ model = PositionalAttribute
- def list(self, filters={}):
- attrs = [StructuralAttribute(self.client, self.corpus, attr) for attr
- in self.client.api.corpus_structural_attributes(
- self.corpus.name)]
- for k, v in filters.items():
- if k == 'part_of':
- attrs = list(filter(lambda x: x.name.startswith(v.name + '_'),
- attrs))
- return attrs
+ def list(self):
+ return [self.prepare_model(self.model._attrs(self.client, self.corpus, attr)) # noqa
+ for attr in self.client.api.corpus_positional_attributes(self.corpus.attrs['api_name'])] # noqa
class StructuralAttribute(Attribute):
- def __init__(self, client, corpus, name):
- super(StructuralAttribute, self).__init__(client, corpus, name)
- self.has_values = \
- client.api.corpus_structural_attribute_has_values(self._name)
+ @staticmethod
+ def _attrs(client, corpus, name):
+ attrs = super(StructuralAttribute, StructuralAttribute)._attrs(client, corpus, name) # noqa
+ attrs['has_values'] = client.api.corpus_structural_attribute_has_values(attrs['api_name']) # noqa
+ return attrs
def cpos_by_id(self, id):
- return self.client.api.cl_struc2cpos(self._name, id)
+ return self.client.api.cl_struc2cpos(self.attrs['api_name'], id)
def ids_by_cpos(self, cpos_list):
- return self.client.api.cl_cpos2struc(self._name, cpos_list)
+ return self.client.api.cl_cpos2struc(self.attrs['api_name'], cpos_list)
def lbound_by_cpos(self, cpos_list):
- return self.client.api.cl_cpos2lbound(self._name, cpos_list)
+ return self.client.api.cl_cpos2lbound(self.attrs['api_name'],
+ cpos_list)
def rbound_by_cpos(self, cpos_list):
- return self.client.api.cl_cpos2rbound(self._name, cpos_list)
+ return self.client.api.cl_cpos2rbound(self.attrs['api_name'],
+ cpos_list)
def values_by_ids(self, id_list):
- return self.client.api.cl_struc2str(self._name, id_list)
+ return self.client.api.cl_struc2str(self.attrs['api_name'], id_list)
def export(self, first_cpos, last_cpos, context=0, expand_lists=False):
first_id, last_id = self.ids_by_cpos([first_cpos, last_cpos])
@@ -150,9 +130,10 @@ class StructuralAttribute(Attribute):
else:
cpos_list_lbound = first_cpos
rc_rbound = \
- self.cpos_by_id(min((last_id + context), (self.size - 1)))[1]
+ self.cpos_by_id(min((last_id + context), (self.attrs['size'] - 1)))[1]
if rc_rbound != last_cpos:
- rc_lbound = min((last_cpos + 1), (self.corpus.size - 1))
+ rc_lbound = min((last_cpos + 1),
+ (self.collection.corpus.attrs['size'] - 1))
rc = (rc_lbound, rc_rbound)
cpos_list_rbound = rc_rbound
else:
@@ -164,5 +145,17 @@ class StructuralAttribute(Attribute):
'rc': list(range(rc[0], (rc[1] + 1)))}
else:
match = {'lc': lc, 'c': c, 'rc': rc}
- lookups = self.corpus.lookups_by_cpos(cpos_list)
+ lookups = self.collection.corpus.lookups_by_cpos(cpos_list)
return {'match': match, **lookups}
+
+
+class StructuralAttributeCollection(AttributeCollection):
+ model = StructuralAttribute
+
+ def list(self, filters={}):
+ attrs = [self.prepare_model(self.model._attrs(self.client, self.corpus, attr)) # noqa
+ for attr in self.client.api.corpus_structural_attributes(self.corpus.attrs['api_name'])] # noqa
+ for k, v in filters.items():
+ if k == 'part_of':
+ attrs = list(filter(lambda x: x.attrs['name'].startswith(v.attrs['name'] + '_'), attrs)) # noqa
+ return attrs
diff --git a/app/corpora/cqi/models/corpora.py b/app/corpora/cqi/models/corpora.py
index e81056ab93df8ef58b37ff2c44f29ec7af2c2429..a12d9b4f8ef54987401014ea36a780aff085e933 100644
--- a/app/corpora/cqi/models/corpora.py
+++ b/app/corpora/cqi/models/corpora.py
@@ -1,30 +1,24 @@
-from .attributes import AttributeCollection
+from .attributes import (AlignmentAttributeCollection,
+ PositionalAttributeCollection,
+ StructuralAttributeCollection)
+from .ressource import Collection, Model
from .subcorpora import SubcorpusCollection
-class CorpusCollection:
- def __init__(self, client):
- self.client = client
+class Corpus(Model):
+ id_attribute = 'api_name'
- def get(self, name):
- return Corpus(self.client, name)
-
- def list(self):
- return [Corpus(self.client, corpus) for corpus in
- self.client.api.corpus_list_coprora()]
-
-
-class Corpus:
- def __init__(self, client, name):
- self.client = client
- self.name = name
- self.size = client.api.cl_attribute_size('{}.word'.format(name))
- # self.info = client.api.corpus_info(name)
- self.charset = client.api.corpus_charset(name)
- # self.full_name = client.api.corpus_full_name(name)
- self.properties = client.api.corpus_properties(name)
- self.attributes = AttributeCollection(client, self)
- self.subcorpora = SubcorpusCollection(client, self)
+ @staticmethod
+ def _attrs(client, name):
+ api_name = name
+ return {'api_name': api_name,
+ 'name': name,
+ 'size': client.api.cl_attribute_size(
+ '{}.word'.format(api_name)),
+ # 'info': client.api.corpus_info(name),
+ 'charset': client.api.corpus_charset(api_name),
+ # 'full_name' = client.api.corpus_full_name(name),
+ 'properties': client.api.corpus_properties(api_name)}
def lookups_by_cpos(self, cpos_list):
cpos_list = list(set(cpos_list))
@@ -33,35 +27,64 @@ class Corpus:
lookups['cpos_lookup'] = {}
for cpos in cpos_list:
lookups['cpos_lookup'][cpos] = {}
- for attr in self.attributes.positional.list():
+ for attr in self.positional_attributes.list():
cpos_attr_values = attr.values_by_cpos(cpos_list)
for i, cpos in enumerate(cpos_list):
- lookups['cpos_lookup'][cpos][attr.name] = cpos_attr_values[i]
- for attr in self.attributes.structural.list():
- if attr.has_values:
+ lookups['cpos_lookup'][cpos][attr.attrs['name']] = cpos_attr_values[i]
+ for attr in self.structural_attributes.list():
+ if attr.attrs['has_values']:
continue
cpos_attr_ids = attr.ids_by_cpos(cpos_list)
for i, cpos in enumerate(cpos_list):
if cpos_attr_ids[i] != -1:
- lookups['cpos_lookup'][cpos][attr.name] = cpos_attr_ids[i]
+ lookups['cpos_lookup'][cpos][attr.attrs['name']] = cpos_attr_ids[i]
occured_attr_ids = list(filter(lambda x: x != -1,
set(cpos_attr_ids)))
if not occured_attr_ids:
continue
subattrs = \
- self.attributes.structural.list(filters={'part_of': attr})
+ self.structural_attributes.list(filters={'part_of': attr})
if not subattrs:
continue
- lookup_name = '{}_lookup'.format(attr.name)
+ lookup_name = '{}_lookup'.format(attr.attrs['name'])
lookups[lookup_name] = {}
for attr_id in occured_attr_ids:
lookups[lookup_name][attr_id] = {}
for subattr in subattrs:
subattr_values = subattr.values_by_ids(occured_attr_ids)
for i, subattr_value in enumerate(subattr_values):
- lookups[lookup_name][occured_attr_ids[i]][subattr.name] = \
+ subattr_name = subattr.attrs['name'][(len(attr.attrs['name']) + 1):]
+ lookups[lookup_name][occured_attr_ids[i]][subattr_name] = \
subattr_value
return lookups
def query(self, query, subcorpus_name='Results'):
- return self.client.api.cqp_query(self.name, subcorpus_name, query)
+ return self.client.api.cqp_query(self.attrs['api_name'],
+ subcorpus_name, query)
+
+ @property
+ def alignment_attributes(self):
+ return AlignmentAttributeCollection(client=self.client, corpus=self)
+
+ @property
+ def positional_attributes(self):
+ return PositionalAttributeCollection(client=self.client, corpus=self)
+
+ @property
+ def structural_attributes(self):
+ return StructuralAttributeCollection(client=self.client, corpus=self)
+
+ @property
+ def subcorpora(self):
+ return SubcorpusCollection(client=self.client, corpus=self)
+
+
+class CorpusCollection(Collection):
+ model = Corpus
+
+ def get(self, corpus_name):
+ return self.prepare_model(self.model._attrs(self.client, corpus_name))
+
+ def list(self):
+ return [self.prepare_model(self.model._attrs(self.client, corpus))
+ for corpus in self.client.api.corpus_list_coprora()]
diff --git a/app/corpora/cqi/models/ressource.py b/app/corpora/cqi/models/ressource.py
new file mode 100644
index 0000000000000000000000000000000000000000..449bc93bf34e791748ddda0a1fdb9ac6399ca799
--- /dev/null
+++ b/app/corpora/cqi/models/ressource.py
@@ -0,0 +1,78 @@
+class Model:
+ """
+ A base class for representing a single object on the server.
+ """
+ id_attribute = 'Id'
+
+ def __init__(self, attrs=None, client=None, collection=None):
+ #: A client pointing at the server that this object is on.
+ self.client = client
+
+ #: The collection that this model is part of.
+ self.collection = collection
+
+ #: The raw representation of this object from the API
+ self.attrs = attrs or {}
+
+ @staticmethod
+ def _attrs(client, key):
+ raise NotImplementedError
+
+ def __repr__(self):
+ return "<{}: {}>".format(self.__class__.__name__, self.id)
+
+ def __eq__(self, other):
+ return isinstance(other, self.__class__) and self.id == other.id
+
+ def __hash__(self):
+ return hash("{}:{}".format(self.__class__.__name__, self.id))
+
+ @property
+ def id(self):
+ """
+ The ID of the object.
+ """
+ return self.attrs.get(self.id_attribute)
+
+ def reload(self):
+ """
+ Load this object from the server again and update ``attrs`` with the
+ new data.
+ """
+ new_model = self.collection.get(self.id)
+ self.attrs = new_model.attrs
+
+
+class Collection:
+ """
+ A base class for representing all objects of a particular type on the
+ server.
+ """
+
+ #: The type of object this collection represents, set by subclasses
+ model = None
+
+ def __init__(self, client=None):
+ #: The client pointing at the server that this collection of objects
+ #: is on.
+ self.client = client
+
+ def list(self):
+ raise NotImplementedError
+
+ def get(self, key):
+ raise NotImplementedError
+
+ def prepare_model(self, attrs):
+ """
+ Create a model from a set of attributes.
+ """
+ if isinstance(attrs, Model):
+ attrs.client = self.client
+ attrs.collection = self
+ return attrs
+ elif isinstance(attrs, dict):
+ return self.model(attrs=attrs, client=self.client, collection=self)
+ else:
+ raise Exception("Can't create {} from {}".format(
+ self.model.__name__, attrs))
diff --git a/app/corpora/cqi/models/subcorpora.py b/app/corpora/cqi/models/subcorpora.py
index 0e774db0ab370d27d88a5ceff5eda69414794cc8..38d254356cfc67d5914167cb0b1498f63d659634 100644
--- a/app/corpora/cqi/models/subcorpora.py
+++ b/app/corpora/cqi/models/subcorpora.py
@@ -1,54 +1,45 @@
+from .ressource import Collection, Model
from ..api.specification import (CONST_FIELD_KEYWORD, CONST_FIELD_MATCH,
CONST_FIELD_MATCHEND, CONST_FIELD_TARGET)
-class SubcorpusCollection:
- def __init__(self, client, corpus):
- self.client = client
- self.corpus = corpus
-
- def get(self, name):
- return Subcorpus(self.client, self.corpus, name)
-
- def list(self):
- return [Subcorpus(self.client, self.corpus, subcorpus) for subcorpus in
- self.client.api.cqp_list_subcorpora(self.corpus.name)]
+class Subcorpus(Model):
+ id_attribute = 'api_name'
-
-class Subcorpus:
- def __init__(self, client, corpus, name):
- self.client = client
- self.corpus = corpus
- self.name = name
- self._name = '{}:{}'.format(corpus.name, name)
- self.fields = {}
- if client.api.cqp_subcorpus_has_field(self._name, CONST_FIELD_MATCH):
- self.fields['match'] = CONST_FIELD_MATCH
- if client.api.cqp_subcorpus_has_field(self._name,
+ @staticmethod
+ def _attrs(client, corpus, name):
+ api_name = '{}:{}'.format(corpus.attrs['api_name'], name)
+ fields = {}
+ if client.api.cqp_subcorpus_has_field(api_name, CONST_FIELD_MATCH):
+ fields['match'] = CONST_FIELD_MATCH
+ if client.api.cqp_subcorpus_has_field(api_name,
CONST_FIELD_MATCHEND):
- self.fields['matchend'] = CONST_FIELD_MATCHEND
- if client.api.cqp_subcorpus_has_field(self._name, CONST_FIELD_TARGET):
- self.fields['target'] = CONST_FIELD_TARGET
- if client.api.cqp_subcorpus_has_field(self._name, CONST_FIELD_KEYWORD):
- self.fields['keyword'] = CONST_FIELD_KEYWORD
- self.size = client.api.cqp_subcorpus_size(self._name)
+ fields['matchend'] = CONST_FIELD_MATCHEND
+ if client.api.cqp_subcorpus_has_field(api_name, CONST_FIELD_TARGET):
+ fields['target'] = CONST_FIELD_TARGET
+ if client.api.cqp_subcorpus_has_field(api_name, CONST_FIELD_KEYWORD):
+ fields['keyword'] = CONST_FIELD_KEYWORD
+ return {'api_name': api_name,
+ 'name': name,
+ 'fields': fields,
+ 'size': client.api.cqp_subcorpus_size(api_name)}
def drop(self):
- return self.client.api.cqp_drop_subcorpus(self._name)
+ return self.client.api.cqp_drop_subcorpus(self.attrs['api_name'])
def dump(self, field, first, last):
- return self.client.api.cqp_dump_subcorpus(self._name, field, first,
- last)
+ return self.client.api.cqp_dump_subcorpus(self.attrs['api_name'],
+ field, first, last)
def export(self, context=25, cutoff=float('inf'), expand_lists=False,
offset=0):
- if self.size == 0:
+ if self.attrs['size'] == 0:
return {"matches": []}
first_match = max(0, offset)
- last_match = min((offset + cutoff - 1), (self.size - 1))
+ last_match = min((offset + cutoff - 1), (self.attrs['size'] - 1))
match_boundaries = \
- zip(self.dump(self.fields['match'], first_match, last_match),
- self.dump(self.fields['matchend'], first_match, last_match))
+ zip(self.dump(self.attrs['fields']['match'], first_match, last_match), # noqa
+ self.dump(self.attrs['fields']['matchend'], first_match, last_match)) # noqa
cpos_list = []
matches = []
for match_start, match_end in match_boundaries:
@@ -64,10 +55,11 @@ class Subcorpus:
cpos_list_lbound = lc_lbound
else:
cpos_list_lbound = match_start
- rc_lbound = min((match_end + 1), (self.corpus.size - 1))
+ rc_lbound = min((match_end + 1),
+ (self.collection.corpus.attrs['size'] - 1))
if rc_lbound != match_end:
rc_rbound = min((match_end + 1 + context),
- (self.corpus.size - 1))
+ (self.collection.corpus.attrs['size'] - 1))
rc = (rc_lbound, rc_rbound)
cpos_list_rbound = rc_rbound
else:
@@ -81,14 +73,30 @@ class Subcorpus:
else:
match = {'lc': lc, 'c': c, 'rc': rc}
matches.append(match)
- lookups = self.corpus.lookups_by_cpos(cpos_list)
+ lookups = self.collection.corpus.lookups_by_cpos(cpos_list)
return {'matches': matches, **lookups}
def fdist_1(self, cutoff, field, attribute):
- return self.client.api.cqp_fdist_1(self._name, cutoff,
+ return self.client.api.cqp_fdist_1(self.attrs['api_name'], cutoff,
field, attribute._name)
def fdist_2(self, cutoff, field_1, attribute_1, field_2, attribute_2):
- return self.client.api.cqp_fdist_2(self._name, cutoff,
+ return self.client.api.cqp_fdist_2(self.attrs['api_name'], cutoff,
field_1, attribute_1._name,
field_2, attribute_2._name)
+
+
+class SubcorpusCollection(Collection):
+ model = Subcorpus
+
+ def __init__(self, client=None, corpus=None):
+ super(SubcorpusCollection, self).__init__(client=client)
+ self.corpus = corpus
+
+ def get(self, subcorpus_name):
+ return self.prepare_model(self.model._attrs(self.client, self.corpus,
+ subcorpus_name))
+
+ def list(self):
+ return [self.prepare_model(self.model._attrs(self.client, self.corpus, subcorpus)) # noqa
+ for subcorpus in self.client.api.cqp_list_subcorpora(self.corpus.attrs['api_name'])] # noqa
diff --git a/app/corpora/cqi/wrapper.py b/app/corpora/cqi/wrapper.py
deleted file mode 100644
index f6c58395bdebc685ca753d05d457370df96fcd75..0000000000000000000000000000000000000000
--- a/app/corpora/cqi/wrapper.py
+++ /dev/null
@@ -1,321 +0,0 @@
-from .api import APIClient
-from .api.specification import CONST_FIELD_MATCH, CONST_FIELD_MATCHEND
-import time
-
-
-class CQiWrapper(APIClient):
- '''
- CQIiWrapper object
-
- High level wrapper that groups and renames some functions of CQiClient
- for ease of use. Also structures recieved data into python dictionaries.
-
- Keyword arguments:
- host -- host IP adress or hostname wher the cqp server is running
- port -- port of the cqp server
- username -- username used to connect to the cqp server
- password -- password of the user to connect to the cqp server
- '''
-
- SUBCORPUS_NAMES = []
-
- def __init__(self, host='127.0.0.1', port=4877, username='anonymous',
- password=''):
- super(CQiWrapper, self).__init__(host, port=port)
- self.username = username
- self.password = password
-
- def connect(self):
- '''
- Connect with CQP server
-
- Connects via socket to the CQP server using the given username and
- password from class initiation.
- '''
- self.ctrl_connect(self.username, self.password)
-
- def __create_attribute_strings(self):
- '''
- Creates all needed attribute strings to query for word, lemma etc. in
- the given corpus.
- For example: CORPUS_NAME.word to query words
- Automaticalle creates strings for all pre defined tags.
- '''
- p_attrs = self.corpus_positional_attributes(self.corpus_name)
- struct_attrs = self.corpus_structural_attributes(self.corpus_name)
- self.attr_strings = {}
- self.attr_strings['positional_attrs'] = {}
- self.attr_strings['struct_attrs'] = {}
- for p_attr in p_attrs:
- self.attr_strings['positional_attrs'][p_attr] = (self.corpus_name
- + '.'
- + p_attr)
- for struct_attr in struct_attrs:
- self.attr_strings['struct_attrs'][struct_attr] = (self.corpus_name
- + '.'
- + struct_attr)
- print(('All positional and '
- 'structural attributes: {}').format(self.attr_strings))
-
- def select_corpus(self, corpus_name):
- '''
- Checks if given copus name exists. If it exists set it as the main
- corpus name used to create the needed query attribute strings like
- CORPUS_NAME.word.
- '''
- if corpus_name in self.corpus_list_coprora():
- self.corpus_name = corpus_name
- self.__create_attribute_strings()
- print('{} does exist.'.format(corpus_name))
- else:
- print('{} does not exist.'.format(corpus_name))
- raise Exception('Given Corpus Name is not in corpora list.')
-
- def disconnect(self):
- '''
- Disconnect from CQP server
-
- Disconnects from the CQP server. Closes used socket after disconnect.
- '''
- self.ctrl_bye()
- print('Disconnected from cqp server.')
-
- def query_subcorpus(self, query, result_subcorpus_name='Query-results'):
- '''
- Create subcorpus
-
- Input query will be used to create a subcorpus holding all cpos match
- positions for that query.
-
- Keyword arguments:
- result_subcorpus_name -- set name of the subcorpus which holds all
- cpos match positions, produced by the query
- query -- query written in cqp query language
- '''
- self.query = query
- self.cqp_query(self.corpus_name, result_subcorpus_name, query)
- self.result_subcorpus = (self.corpus_name
- + ':'
- + result_subcorpus_name)
- self.SUBCORPUS_NAMES.append(self.result_subcorpus)
- self.match_count = self.cqp_subcorpus_size(self.result_subcorpus)
- print('Nr of all matches is: {}'.format(self.match_count))
-
- def show_subcorpora(self):
- '''
- Show all subcorpora currently saved by the cqp server.
- '''
- return self.cqp_list_subcorpora(self.corpus_name)
-
- def show_query_results(self,
- context_len=10,
- result_len=1000,
- result_offset=0):
- '''
- Show query results
-
- Shows the actual matched strings produce by the query. Uses the cpos
- match indexes to grab those strings. saves them into an orderd
- dictionary. Also saves coresponding tags, lemmas and context. Gets those
- informations using the corresponding cpos.
-
- Keyword arguments:
- context_len -- defines how many words before and after a match will be
- shown (default 10)
- result_len -- defines for how many matches all informations like lemma
- and POS are being grabbed
- result_offset -- defines the offset of the matches being requested. If
- the offset is 100 informations for matches 100 to result_len are being
- grabbed
- '''
- t0 = time.time()
- self.context_len = context_len
- self.corpus_max_len = self.cl_attribute_size(
- self.attr_strings['positional_attrs']['word']
- )
- self.nr_matches = min(result_len, self.match_count)
- if self.match_count == 0:
- print('Query resulted in 0 matches.')
- self.results = {'code': 0,
- 'result': {'matches': [],
- 'match_count': self.match_count,
- 'cpos_lookup': {},
- 'text_lookup': {}}
- }
- return self.results
- else:
- # Get match cpos boundries
- # match_boundries shows the start and end cpos of one match as a
- # pair of cpositions
- # [(1355, 1357), (1477, 1479)] Example for two boundry pairs
- offset_start = 0 if result_offset == 0 else result_offset
- print('Offset start is: {}'.format(offset_start))
- offset_end = min((self.nr_matches + result_offset - 1), self.match_count - 1)
- print('Offset end is: {}'.format(offset_end))
- match_boundaries = zip(self.cqp_dump_subcorpus(self.result_subcorpus,
- CONST_FIELD_MATCH,
- offset_start,
- offset_end),
- self.cqp_dump_subcorpus(self.result_subcorpus,
- CONST_FIELD_MATCHEND,
- offset_start,
- offset_end))
-
- # Generate all cpos between match boundries including start and end
- # boundries.
- # Also generate cpos for left and right context.
- # Save those cpos into dict as lists for the keys 'lc', 'hit' and 'rc'
- # Also collect all cpos together in one list for the final request of
- # all cpos informations
- all_matches = []
- all_cpos = []
- for start, end in match_boundaries:
- end += 1
- lc_cpos = list(range(max([0, start - self.context_len]), start))
- lc = {'lc': lc_cpos}
- match_cpos = list(range(start, end))
- match = {'hit': match_cpos}
- rc_cpos = list(range(end, min([self.corpus_max_len,
- end + self.context_len])))
- rc = {'rc': rc_cpos}
- lc.update(match)
- lc.update(rc)
- all_cpos.extend(lc_cpos + match_cpos + rc_cpos)
- all_matches.append(lc)
-
- all_cpos = list(set(all_cpos)) # get rid of cpos duplicates
- len_all_cpos = len(all_cpos)
- t1 = time.time()
- t_total = t1 - t0
- print('Time to create all CPOS for query: {}'.format(t_total))
- print('Requesting {} CPOS with one query.'.format(len_all_cpos))
-
- # Get cpos informations like CORPUS_NAME.word or CORPUS_NAME.lemma for
- # all cpos entries in all_cpos_list
- # Also saves these informations into self.results dict
- t2 = time.time()
- all_cpos_infos, text_lookup = self.get_cpos_infos(all_cpos)
- t3 = time.time()
- t_final = t3 - t2
- print('Got infos for {} CPOS in {} seconds:'.format(len_all_cpos,
- t_final))
- self.results = {'code': 0,
- 'result': {'matches': all_matches,
- 'match_count': self.match_count,
- 'cpos_lookup': all_cpos_infos,
- 'text_lookup': text_lookup}
- }
- return self.results
-
- def get_cpos_infos(self, all_cpos):
- '''
- Get cpos informations like CORPUS_NAME.word or CORPUS_NAME.lemma for
- all cpos entries specified in the parameter all_cpos.
- '''
- # Get all positional attribute informations
- cpos_infos = {}
- for p_attr_key in self.attr_strings['positional_attrs'].keys():
- match_strs = self.cl_cpos2str(self.attr_strings['positional_attrs'][p_attr_key], all_cpos)
- cpos_infos[p_attr_key] = match_strs
-
- # Get all strucutural attribute informations
- tmp_info = {}
- structs_to_check = []
- for struct_attr_key in self.attr_strings['struct_attrs'].keys():
- key = self.attr_strings['struct_attrs'][struct_attr_key]
- has_value = self.corpus_structural_attribute_has_values(key)
- struct_ids = self.cl_cpos2struc(key, all_cpos)
- if has_value is False: # Get IDs of strucutural elements without values (this means get IDs of XML tags. Struct elements only have values if they are XML attributes)
- tmp_info[struct_attr_key] = []
- for id in struct_ids:
- tmp_info[struct_attr_key].append(id)
- else:
- structs_to_check.append({key: struct_attr_key})
- print('Structs to check: {}'.format(structs_to_check))
- struct_attr_values = list(tmp_info.values())
- # print('Struct attr value list: {}'.format(struct_attr_values))
- struct_attr_keys = list(tmp_info.keys())
- # print('Struct attr key list: {}'.format(struct_attr_keys))
-
- # Build textlookup dictionary
- text_lookup_ids = list(set(struct_attr_values[0])) # every CPOS is associated with one text id. A set is build to only gather text_lookup informations for every unique text id
- text_lookup = {} # final dict containing all info of one text identified by its id
- for d in structs_to_check:
- s_key, s_value = zip(*d.items())
- print('dict entries: {}: {}'.format(s_key, s_value))
- s_value = s_value[0].split('_', 1)[-1]
- print('S_VALUE: {}'.format(s_value))
- struct_values = self.cl_struc2str(s_key[0], text_lookup_ids)
- print('Extracted Value with key {}: {}'.format(s_key[0], struct_values))
- zipped = dict(zip(text_lookup_ids, struct_values))
- for zip_key, zip_value in zipped.items():
- print('Text id as key is: {}'.format(zip_key))
- print('Value of this text is: {}'.format(zip_value))
- check = text_lookup.get(zip_key)
- print('check: {}'.format(check))
- if check is None:
- text_lookup[zip_key] = {s_value: zip_value}
- else:
- text_lookup[zip_key].update({s_value: zip_value})
-
- # zip keys and values together
- attr_values_list = []
- attr_keys_list = []
- for key in cpos_infos.keys():
- attr_values_list.append(cpos_infos[key])
- attr_keys_list.append(key)
- attr_keys_list.extend(struct_attr_keys)
- attr_values_list.extend(struct_attr_values)
- joined_cpos_infos = zip(all_cpos, *attr_values_list)
- dict_cpos_infos = {}
- for info in joined_cpos_infos:
- dict_cpos_infos[info[0]] = dict(zip(attr_keys_list, info[1:]))
- return dict_cpos_infos, text_lookup
-
- def get_sentences(self,
- match_cpos_list,
- get_surrounding_s=False,
- l_r_s_context_additional_len=1):
- '''
- Get sentence informations for one match also set if and how much left
- right context sentences should be grabbed surrounding the given CPOS.
- '''
- t0 = time.time()
- key = self.corpus_name + '.s'
- first_cpos, last_cpos = match_cpos_list[0], match_cpos_list[-1]
- context_sentences = {}
- s_ids = self.cl_cpos2struc(key, [first_cpos, last_cpos])
- print('s id match: {}'.format(s_ids))
- for s_id in s_ids:
- s_start, s_end = self.cl_struc2cpos(key, s_id)
- s_cpos = list(range(s_start, s_end + 1))
- context_sentences[s_id] = s_cpos
- if get_surrounding_s:
- max_s_id = self.cl_attribute_size(key) - 1
- print('max sid: {}'.format(max_s_id))
- additional_s_ids = []
- additional_s = list(range(max(s_ids[0]
- - l_r_s_context_additional_len,
- 0),
- min(s_ids[-1]
- + l_r_s_context_additional_len,
- max_s_id) + 1))
- additional_s_ids.extend(additional_s)
- for s_id in additional_s_ids:
- print('s id additional: {}'.format(s_id))
- s_start, s_end = self.cl_struc2cpos(key, s_id)
- s_cpos = list(range(s_start, s_end + 1))
- context_sentences[s_id] = s_cpos
- all_cpos = []
- for key in context_sentences.keys():
- all_cpos.extend(context_sentences[key])
- all_cpos = list(set(all_cpos))
- all_cpos_infos, text_lookup = self.get_cpos_infos(all_cpos)
- t1 = time.time()
- t_total = t1 - t0
- print('Got all sentences informations in {} seconds'. format(t_total))
- match_context = {'context_s_cpos': context_sentences,
- 'cpos_lookup': all_cpos_infos,
- 'text_lookup': text_lookup,
- 'match_cpos_list': match_cpos_list}
- return match_context
diff --git a/app/corpora/events.py b/app/corpora/events.py
index 73d189f9d1ef9a6e437314906c337825628ab0d2..cf9c791ceb997b0429df3b5e3c41c59e25db6ec6 100644
--- a/app/corpora/events.py
+++ b/app/corpora/events.py
@@ -67,23 +67,23 @@ def corpus_analysis_query(query):
socketio.emit('corpus_analysis_query', response, room=request.sid)
return
response = {'code': 200, 'desc': None, 'msg': 'OK',
- 'payload': {**query_status, 'match_count': results.size}}
+ 'payload': {**query_status, 'match_count': results.attrs['size']}}
socketio.emit('corpus_analysis_query', response, room=request.sid)
chunk_size = 100
chunk_start = 0
context = 100
progress = 0
client.status = 'running'
- while chunk_start <= results.size:
+ while chunk_start <= results.attrs['size']:
if client.status == 'abort':
break
chunk = results.export(context=context, cutoff=chunk_size,
expand_lists=False, offset=chunk_start)
chunk['cpos_ranges'] = True
- if (results.size == 0):
+ if (results.attrs['size'] == 0):
progress = 100
else:
- progress = ((chunk_start + chunk_size) / results.size) * 100
+ progress = ((chunk_start + chunk_size) / results.attrs['size']) * 100
progress = min(100, int(math.ceil(progress)))
response = {'code': 200, 'desc': None, 'msg': 'OK',
'payload': {'chunk': chunk, 'progress': progress}}
@@ -104,12 +104,12 @@ def corpus_analysis_inspect_match(payload):
return
# Get more context for given match CPOS
corpus = client.corpora.get('CORPUS')
- s = corpus.attributes.structural.get('s')
+ s = corpus.structural_attributes.get('s')
match_context = s.export(payload['first_cpos'], payload['last_cpos'],
context=3, expand_lists=False)
match_context['cpos_ranges'] = True
- socketio.emit('match_context',
- {'payload': match_context}, room=request.sid)
+ socketio.emit('match_context', {'payload': match_context},
+ room=request.sid)
def corpus_analysis_session_handler(app, corpus_id, user_id, session_id):
diff --git a/app/static/js/nopaque.lists.js b/app/static/js/nopaque.lists.js
index 33ba4587b18d12d5d79ce29c805b804afa3762d6..f722e1e45eee848b0b09e42f66b25c5463160e48 100644
--- a/app/static/js/nopaque.lists.js
+++ b/app/static/js/nopaque.lists.js
@@ -381,7 +381,7 @@ class ResultsList extends List {
token = chunk["cpos_lookup"][cpos];
hitCellElement.insertAdjacentHTML("beforeend", `<span class="token" data-cpos="${cpos}">${token["word"]} </span>`);
// get text titles of every hit cpos token
- textTitles.add(chunk["text_lookup"][token["text"]]["text_title"]);
+ textTitles.add(chunk["text_lookup"][token["text"]]["title"]);
// add button to trigger more context to every match td
var inspectBtn = document.createElement("a");
inspectBtn.setAttribute("class", "btn-floating btn-flat waves-effect waves-light grey right inspect disabled");