Merge pull request #42 from INCATools/search

Refactored search

Makefile

@@ -5,7 +5,7 @@ test:
 
 ## Compiled
 
-MODELS = ontology_metadata  obograph  validation_datamodel summary_statistics_datamodel lexical_index mapping_rules_datamodel text_annotator search_results oxo taxon_constraints similarity
+MODELS = ontology_metadata  obograph  validation_datamodel summary_statistics_datamodel lexical_index mapping_rules_datamodel text_annotator search_results oxo taxon_constraints similarity search_datamodel
 
 pyclasses: $(patsubst %, src/oaklib/datamodels/%.py, $(MODELS))
 jsonschema: $(patsubst %, src/oaklib/datamodels/%.schema.json, $(MODELS))

docs/datamodels/funowl/index.md

@@ -1,3 +1,5 @@
+.. funowl:
+
 # FunOwl
 
 FunOwl provided an OWL axiom-oriented datamodel, contrasting from a more graph-oriented data scientist oriented data model

docs/datamodels/index.rst

@@ -20,6 +20,7 @@ not need this.
    :caption: Contents:
 
    ontology-metadata/index
+   search/index
    obograph/index
    sssom/index
    funowl/index
@@ -31,3 +32,4 @@ not need this.
    text-annotator/index
 
 
+

docs/intro/tutorial04.rst

@@ -18,6 +18,8 @@ To illustrate this we will first introduce the :ref:`OboGraphInterface` which us
 OboGraphs
 ---------
 
+See `OboGraph Datamodel <https://incatools.github.io/ontology-access-kit/datamodels/obograph/index.html>`_
+
 In this datamodel, an ontology is conceived of as a Graph, with a graph consisting of
 
 - nodes

docs/intro/tutorial05.rst

@@ -15,7 +15,7 @@ Command Line Usage
 
     runoak -i obolibrary:fbbt.obo viz FBbt:00004751 -p i,p -o wing-vein.png
 
-Then open wing-vein.png
+Then open the file wing-vein.png using a tool such as Preview, or in a web browser.
 
 If you omit the --output option then the png will be written to a temp file and immediately opened:
 
@@ -26,7 +26,9 @@ If you omit the --output option then the png will be written to a temp file and
 How this works
 ---------------
 
- - First,
+ - First, the :ref:`OboGraphInterface` is invoked to query all ancestors of the term
+ - in this case the implementation is :ref:`pronto` plus graph walking code
+ - after an obograph is generated this is passed to the obographviz library
 
 Programmatic usage
 ---------------

docs/intro/tutorial06.rst

@@ -1,13 +1,18 @@
-Part 6: Working With OWO
+Part 6: Working With OWL
 =====================
 
+.. warning ::
+
+   this functionality is not available until v0.3.0
+   we are waiting for one rdflib PR (https://github.com/RDFLib/rdflib/pull/1686) to be merged...
+
 OAK comes bundled with the `funowl <https://github.com/hsolbrig/funowl/>`_ library
 
-TODO: funowl is not yet bundled, we are waiting for one rdflib PR (https://github.com/RDFLib/rdflib/pull/1686) to be merged...
 
 OWL Datamodel
 --------------
 
+See :ref:`funowl`
 
 OwlInterface
 ------------

docs/intro/tutorial07.rst

@@ -1,2 +1,17 @@
 Part 7: SQLite files
-=============
+=============
+
+Download a SQLite file
+----------------
+
+Query
+------
+
+Building your own SQLite files
+-------------------
+
+Python
+------
+
+Other RDBMSs
+------------

pyproject.toml

@@ -14,7 +14,7 @@ SPARQLWrapper = "^2.0.0"
 SQLAlchemy = "^1.4.32"
 linkml-runtime = "^1.2.3"
 networkx = "^2.7.1"
-sssom = "^0.3.7"
+sssom = "^0.3.8"
 ratelimit = "^2.2.1"
 appdirs = "^1.4.4"
 

src/oaklib/cli.py

@@ -16,6 +16,7 @@
 import click
 import rdflib
 from linkml_runtime.dumpers import yaml_dumper, json_dumper
+from oaklib.datamodels.search import create_search_configuration
 from oaklib.datamodels.validation_datamodel import ValidationConfiguration
 from oaklib.implementations.aggregator.aggregator_implementation import AggregatorImplementation
 from oaklib.implementations.sqldb.sql_implementation import SqlImplementation
@@ -179,7 +180,8 @@ def search(terms, output: str):
     impl = settings.impl
     if isinstance(impl, SearchInterface):
         for t in terms:
-            for curie_it in chunk(impl.basic_search(t)):
+            cfg = create_search_configuration(t)
+            for curie_it in chunk(impl.basic_search(cfg.search_terms[0], config=cfg)):
                 logging.info('** Next chunk:')
                 for curie, label in impl.get_labels_for_curies(curie_it):
                     print(f'{curie} ! {label}')

src/oaklib/datamodels/search.py

@@ -0,0 +1,75 @@
+from dataclasses import dataclass
+from typing import List
+
+from deprecated.classic import deprecated
+from oaklib.datamodels.search_datamodel import SearchBaseConfiguration, SearchProperty, SearchTermSyntax
+from oaklib.datamodels.vocabulary import LABEL_PREDICATE, SYNONYM_PREDICATES
+from oaklib.types import PRED_CURIE
+
+DEFAULT_SEARCH_PROPERTIES = [SearchProperty.LABEL, SearchProperty.ALIAS]
+
+def create_search_configuration(term: str) -> "SearchConfiguration":
+
+    if len(term) > 1:
+        prop = term[0]
+        code = term[1]
+        rest = term[2:]
+        if code == '~':
+            cfg = SearchBaseConfiguration([rest],
+                                          is_partial=True)
+            syntax = SearchTermSyntax.PLAINTEXT
+        elif code == '/':
+            cfg = SearchBaseConfiguration([rest],
+                                          syntax=SearchTermSyntax.REGULAR_EXPRESSION)
+        elif code == '=':
+            cfg = SearchBaseConfiguration([rest],
+                                          is_partial=False)
+        elif code == '^':
+            cfg = SearchBaseConfiguration([rest],
+                                          syntax=SearchTermSyntax.STARTS_WITH)
+        else:
+            cfg = SearchBaseConfiguration([term], properties=DEFAULT_SEARCH_PROPERTIES)
+            prop = None
+        if prop:
+            if prop == 't':
+                props = [SearchProperty.LABEL, SearchProperty.ALIAS]
+            elif prop == '.':
+                props = [SearchProperty.ANYTHING]
+            elif prop == 'l':
+                props = [SearchProperty.LABEL]
+            else:
+                raise ValueError(f'Unknown property code: {prop}')
+            cfg.properties = [SearchProperty(p) for p in props]
+        return cfg
+    else:
+        return SearchConfiguration([term], properties=DEFAULT_SEARCH_PROPERTIES)
+
+def search_properties_to_predicates(props: List[SearchProperty]) -> List[PRED_CURIE]:
+    preds = set()
+    for p in props:
+        if p == SearchProperty(SearchProperty.LABEL):
+            preds.add(LABEL_PREDICATE)
+        elif p == SearchProperty(SearchProperty.ALIAS):
+            preds.update(SYNONYM_PREDICATES + [LABEL_PREDICATE])
+        else:
+            raise ValueError(p)
+    return list(preds)
+
+@dataclass
+class SearchConfiguration(SearchBaseConfiguration):
+    """
+    Parameters for altering behavior of search
+
+    .. note ::
+
+        many of these parameters are not yet implemented
+    """
+
+    @deprecated()
+    def use_label_only(self) -> "SearchConfiguration":
+        self.include_label = False
+        self.include_id = False
+        self.include_definition = False
+        self.include_aliases = False
+        self.properties = [SearchProperty.LABEL]
+        return self