Remove preferred reference

README.rst

@@ -97,7 +97,7 @@ Variables of the methods are defined as follows
 
     * `Reference <http://github.com/openmicroanalysis/pyxray/blob/master/pyxray/descriptor.py>`_ object
     * BibTeX key of a reference
-    * ``None``, the default reference will be used or the first reference found
+    * ``None``, the newest reference will be returned
 
 Element properties
 ------------------

pyxray/base.py

@@ -55,38 +55,6 @@ class NotFound(Exception):
 
 
 class _DatabaseMixin(metaclass=abc.ABCMeta):
-    @abc.abstractmethod
-    def get_preferred_references(self):
-        """
-        Return the BibTeX keys of the preferred references.
-
-        If no reference is specified when calling a method, the value for the first preferred reference is returned.
-        If no preferred reference, the first value is returned.
-
-        :return: preferred references
-        :rtype: :class:`tuple`
-        """
-        raise NotImplementedError
-
-    @abc.abstractmethod
-    @formatdoc(**_docextras)
-    def add_preferred_reference(self, reference):
-        """
-        Adds a preferred reference.
-
-        :arg reference: :class:`Reference` or its BibTeX key
-
-        {exception}
-        """
-        raise NotImplementedError
-
-    @abc.abstractmethod
-    def clear_preferred_references(self):
-        """
-        Clear all added preferred references.
-        """
-        raise NotImplementedError
-
     @abc.abstractmethod
     @formatdoc(**_docextras)
     def element(self, element):  # pragma: no cover

pyxray/data.py

@@ -3,9 +3,6 @@
 """
 
 __all__ = [
-    "get_preferred_references",
-    "add_preferred_reference",
-    "clear_preferred_references",
     "element",
     "element_atomic_number",
     "element_symbol",
@@ -48,15 +45,6 @@
 
 
 class _EmptyDatabase(_DatabaseMixin):
-    def get_preferred_references(self):
-        return ()
-
-    def add_preferred_reference(self, reference):
-        pass
-
-    def clear_preferred_references(self):
-        pass
-
     def element(self, element):  # pragma: no cover
         raise NotFound
 
@@ -169,9 +157,6 @@ def _init_sql_database():
     logger.error("No SQL database found")
     database = _EmptyDatabase()
 
-get_preferred_references = database.get_preferred_references
-add_preferred_reference = database.add_preferred_reference
-clear_preferred_references = database.clear_preferred_references
 element = database.element
 element_atomic_number = database.element_atomic_number
 element_symbol = database.element_symbol

pyxray/descriptor.py

@@ -352,7 +352,7 @@ def __repr__(self):
 class Reference:
     bibtexkey: str
     author: str = None
-    year: str = None
+    year: int = None
     title: str = None
     type: str = None
     booktitle: str = None

pyxray/sql/data.py

@@ -18,10 +18,12 @@
 
 
 class StatementBuilder:
-    def __init__(self):
+    def __init__(self, distinct=False):
+        self._distinct = distinct
         self._columns = []
         self._joins = {}
         self._clauses = []
+        self._orderbys = []
 
     def add_column(self, column):
         self._columns.append(column)
@@ -34,9 +36,15 @@ def add_join(self, left, right, onclause):
     def add_clause(self, clause):
         self._clauses.append(clause)
 
+    def add_orderby(self, column, ascending=True):
+        self._orderbys.append((column, ascending))
+
     def build(self):
         statement = sqlalchemy.sql.select(self._columns)
 
+        if self._distinct:
+            statement = statement.distinct()
+
         if self._joins:
             # Joins have to be nested to work with sqlalchemy
             # E.g.
@@ -54,13 +62,22 @@ def build(self):
 
             statement = statement.select_from(finaljoin)
 
-        return statement.where(sqlalchemy.sql.and_(*self._clauses))
+        if self._clauses:
+            statement = statement.where(sqlalchemy.sql.and_(*self._clauses))
+
+        if self._orderbys:
+            orderbys = [
+                sqlalchemy.asc(column) if ascending else sqlalchemy.desc(column)
+                for column, ascending in self._orderbys
+            ]
+            statement = statement.order_by(*orderbys)
+
+        return statement
 
 
 class SqlDatabase(_DatabaseMixin, SqlBase):
     def __init__(self, engine):
         super().__init__(engine)
-        self._preferred_references = []
 
     def _expand_atomic_subshell(self, atomic_subshell):
         if (
@@ -280,10 +297,10 @@ def _update_reference(self, builder, table, reference, column="reference_id"):
             reference = reference.bibtexkey
 
         table_reference = self.require_table(descriptor.Reference)
-        builder.add_column(table_reference.c["bibtexkey"])
         builder.add_join(
             table, table_reference, table.c[column] == table_reference.c["id"]
         )
+        builder.add_orderby(table_reference.c["year"], ascending=False)  # Newest first
 
         if reference:
             builder.add_clause(table_reference.c["bibtexkey"] == reference)
@@ -308,38 +325,22 @@ def _update_notation(self, builder, table, notation):
         )
         builder.add_clause(table_notation.c["key"] == notation)
 
-    def _execute(self, builder, remove_bibtexkey_column=True):
+    def _execute(self, builder):
         statement = builder.build()
         logger.debug(statement.compile())
 
         # Execute
         with self.engine.connect() as conn:
-            rows = conn.execute(statement).fetchall()
-            if not rows:
+            row = conn.execute(statement).first()
+            if not row:
                 raise NotFound
 
-            # Only one row, no need to check for the preferred references
-            elif len(rows) == 1:
-                row = rows[0]
-
-            else:
-                dictrows = dict((dict(row).get("bibtexkey", None), row) for row in rows)
-                row = rows[0]  # fall back
-                for bibtexkey in self._preferred_references:
-                    if bibtexkey in dictrows:
-                        row = dictrows[bibtexkey]
-                        break
-
-            row = dict(row)
-            if remove_bibtexkey_column:
-                row.pop("bibtexkey", None)
-
             if len(row) == 1:
-                return row.popitem()[1]
+                return row[0]
             else:
-                return row.values()
+                return row
 
-    def _execute_many(self, builder, remove_bibtexkey_column=True):
+    def _execute_many(self, builder):
         statement = builder.build()
         logger.debug(statement.compile())
 
@@ -349,35 +350,7 @@ def _execute_many(self, builder, remove_bibtexkey_column=True):
             if not rows:
                 raise NotFound
 
-            outrows = []
-            for row in rows:
-                row = dict(row)
-                if remove_bibtexkey_column:
-                    row.pop("bibtexkey", None)
-                outrows.append(list(row.values()))
-
-            return outrows
-
-    def get_preferred_references(self):
-        return tuple(self._preferred_references)
-
-    def add_preferred_reference(self, reference):
-        if isinstance(reference, descriptor.Reference):
-            reference = reference.bibtexkey
-        if reference in self._preferred_references:
-            return
-
-        table = self.require_table(descriptor.Reference)
-
-        builder = StatementBuilder()
-        self._update_reference(builder, table, reference, "id")
-
-        bibtexkey = self._execute(builder, remove_bibtexkey_column=False)
-
-        self._preferred_references.append(bibtexkey)
-
-    def clear_preferred_references(self):
-        self._preferred_references.clear()
+            return rows
 
     def element(self, element):
         table = self.require_table(descriptor.Element)
@@ -443,7 +416,7 @@ def element_xray_transitions(self, element, xray_transition=None, reference=None
         table_xray = self.require_table(descriptor.XrayTransition)
         table_probability = self.require_table(property.XrayTransitionProbability)
 
-        builder = StatementBuilder()
+        builder = StatementBuilder(distinct=True)
         builder.add_column(table_xray.c["source_principal_quantum_number"])
         builder.add_column(table_xray.c["source_azimuthal_quantum_number"])
         builder.add_column(table_xray.c["source_total_angular_momentum_nominator"])
@@ -479,7 +452,7 @@ def element_xray_transitions(self, element, xray_transition=None, reference=None
             table_relative_weight = self.require_table(
                 property.XrayTransitionRelativeWeight
             )
-            builder = StatementBuilder()
+            builder = StatementBuilder(distinct=True)
             builder.add_column(table_xray.c["source_principal_quantum_number"])
             builder.add_column(table_xray.c["source_azimuthal_quantum_number"])
             builder.add_column(table_xray.c["source_total_angular_momentum_nominator"])

tests/sql/conftest.py

@@ -20,8 +20,8 @@
 
 class MockParser(_Parser):
     def __iter__(self):
-        reference = descriptor.Reference("lee1966")
-        reference2 = descriptor.Reference("doe2016")
+        reference = descriptor.Reference("lee1966", year=1966)
+        reference2 = descriptor.Reference("doe2016", year=2016)
         element = descriptor.Element(118)
         atomic_shell = descriptor.AtomicShell(1)
         transition = descriptor.XrayTransition(L3, K)

tests/sql/test_data.py

@@ -2,10 +2,6 @@
 """ """
 
 # Standard library modules.
-import os
-import tempfile
-import shutil
-import sqlite3
 
 # Third party modules.
 import pytest
@@ -31,22 +27,6 @@ def database_real(tmp_path):
     return pyxray.data.database
 
 
-def test_add_preferred_reference(database):
-    database.clear_preferred_references()
-    database.add_preferred_reference("lee1966")
-
-    assert len(database.get_preferred_references()) == 1
-    assert "lee1966" in database.get_preferred_references()
-
-    database.clear_preferred_references()
-    assert len(database.get_preferred_references()) == 0
-
-
-def test_add_preferred_reference_not_found(database):
-    with pytest.raises(NotFound):
-        database.add_preferred_reference("foo")
-
-
 @pytest.mark.parametrize("element", [118, "Vi", "Vibranium"])
 def test_element(database, element):
     assert database.element(element) == descriptor.Element(118)
@@ -108,7 +88,8 @@ def test_element_name_notfound_wrong_reference(database):
 
 
 def test_element_atomic_weight_no_reference(database):
-    assert database.element_atomic_weight(118) == pytest.approx(999.1, abs=1e-2)
+    # doe2016 (111.1) takes precedence over lee1966 (999.1) because it is newer
+    assert database.element_atomic_weight(118) == pytest.approx(111.1, abs=1e-2)
 
 
 def test_element_atomic_weight_lee1966(database):
@@ -123,18 +104,6 @@ def test_element_atomic_weight_doe2016(database):
     )
 
 
-def test_element_atomic_weight_preferred_reference(database):
-    database.clear_preferred_references()
-    database.add_preferred_reference("doe2016")
-    assert database.element_atomic_weight(118) == pytest.approx(111.1, abs=1e-2)
-
-    database.clear_preferred_references()
-    database.add_preferred_reference("lee1966")
-    assert database.element_atomic_weight(118) == pytest.approx(999.1, abs=1e-2)
-
-    database.clear_preferred_references()
-
-
 @pytest.mark.parametrize("element", [118, "Vi", "Vibranium"])
 def test_element_mass_density_kg_per_m3(database, element):
     assert database.element_mass_density_kg_per_m3(element) == pytest.approx(