def _write(self, atoms, key_value_pairs, data):
Database._write(self, atoms, key_value_pairs, data)
bigdct = {}
ids = []
nextid = 1
if (isinstance(self.filename, basestring) and
os.path.isfile(self.filename)):
try:
bigdct, ids, nextid = self._read_json()
except (SyntaxError, ValueError):
pass
if isinstance(atoms, AtomsRow):
row = atoms
unique_id = row.unique_id
for id in ids:
if bigdct[id]['unique_id'] == unique_id:
break
else:
id = None
mtime = now()
else:
row = AtomsRow(atoms)
row.ctime = mtime = now()
row.user = os.getenv('USER')
id = None
dct = {}
for key in row.__dict__:
if key[0] == '_' or key in row._keys or key == 'id':
continue
dct[key] = row[key]
dct['mtime'] = mtime
kvp = key_value_pairs or row.key_value_pairs
if kvp:
dct['key_value_pairs'] = kvp
data = data or row.get('data')
if data:
dct['data'] = data
constraints = row.get('constraints')
if constraints:
dct['constraints'] = constraints
if id is None:
id = nextid
ids.append(id)
nextid += 1
bigdct[id] = dct
self._write_json(bigdct, ids, nextid)
return id
def _write(self, atoms, key_value_pairs, data):
Database._write(self, atoms, key_value_pairs, data)
bigdct = {}
ids = []
nextid = 1
if (isinstance(self.filename, basestring)
and os.path.isfile(self.filename)):
try:
bigdct, ids, nextid = self._read_json()
except (SyntaxError, ValueError):
pass
if isinstance(atoms, AtomsRow):
row = atoms
unique_id = row.unique_id
for id in ids:
if bigdct[id]['unique_id'] == unique_id:
break
else:
id = None
mtime = now()
else:
row = AtomsRow(atoms)
row.ctime = mtime = now()
row.user = os.getenv('USER')
id = None
dct = {}
for key in row.__dict__:
if key[0] == '_' or key in row._keys or key == 'id':
continue
dct[key] = row[key]
dct['mtime'] = mtime
kvp = key_value_pairs or row.key_value_pairs
if kvp:
dct['key_value_pairs'] = kvp
data = data or row.get('data')
if data:
dct['data'] = data
constraints = row.get('constraints')
if constraints:
dct['constraints'] = constraints
if id is None:
id = nextid
ids.append(id)
nextid += 1
bigdct[id] = dct
self._write_json(bigdct, ids, nextid)
return id
def _write(self, atoms, key_value_pairs, data, id):
Database._write(self, atoms, key_value_pairs, data)
bigdct = {}
ids = []
nextid = 1
if (isinstance(self.filename, str) and
os.path.isfile(self.filename)):
try:
bigdct, ids, nextid = self._read_json()
except (SyntaxError, ValueError):
pass
mtime = now()
if isinstance(atoms, AtomsRow):
row = atoms
else:
row = AtomsRow(atoms)
row.ctime = mtime
row.user = os.getenv('USER')
dct = {}
for key in row.__dict__:
if key[0] == '_' or key in row._keys or key == 'id':
continue
dct[key] = row[key]
dct['mtime'] = mtime
if key_value_pairs:
dct['key_value_pairs'] = key_value_pairs
if data:
dct['data'] = data
constraints = row.get('constraints')
if constraints:
dct['constraints'] = constraints
if id is None:
id = nextid
ids.append(id)
nextid += 1
else:
assert id in bigdct
bigdct[id] = dct
self._write_json(bigdct, ids, nextid)
return id
def add_project(db: Database) -> None:
"""Add database to projects with name 'default'."""
all_keys: Set[str] = set()
for row in db.select(columns=['key_value_pairs'], include_data=False):
all_keys.update(row._keys)
key_descriptions = {key: (key, '', '') for key in all_keys}
meta: Dict[str, Any] = db.metadata
if 'key_descriptions' in meta:
key_descriptions.update(meta['key_descriptions'])
default_columns = meta.get('default_columns')
if default_columns is None:
default_columns = all_columns[:]
projects['default'] = {
'name': 'default',
'title': meta.get('title', ''),
'uid_key': 'id',
'key_descriptions': create_key_descriptions(key_descriptions),
'database': db,
'row_to_dict_function': row_to_dict,
'handle_query_function': handle_query,
'default_columns': default_columns,
'search_template': 'ase/db/templates/search.html',
'row_template': 'ase/db/templates/row.html',
'table_template': 'ase/db/templates/table.html'
}
def create_table(self, db: Database, uid_key: str,
keys: List[str]) -> Table:
query = self.query
if self.nrows is None:
try:
self.nrows = db.count(query)
except (ValueError, KeyError) as e:
error = ', '.join(['Bad query'] + list(e.args))
from flask import flash
flash(error)
query = 'id=0' # this will return no rows
self.nrows = 0
table = Table(db, uid_key)
table.select(query,
self.columns,
self.sort,
self.limit,
offset=self.page * self.limit,
show_empty_columns=True)
table.format()
assert self.columns is not None
table.addcolumns = sorted(column for column in all_columns + keys
if column not in self.columns)
return table
def _write(self, atoms, keywords, key_value_pairs, data):
Database._write(self, atoms, keywords, key_value_pairs, data)
bigdct = {}
ids = []
nextid = 1
if isinstance(self.filename, str) and os.path.isfile(self.filename):
try:
bigdct, ids, nextid = self._read_json()
except (SyntaxError, ValueError):
pass
if isinstance(atoms, dict):
dct = dict((key, atoms[key])
for key in reserved_keys
if key in atoms and key != 'id')
unique_id = dct['unique_id']
for id in ids:
if bigdct[id]['unique_id'] == unique_id:
break
else:
id = None
dct['mtime'] = now()
else:
dct = self.collect_data(atoms)
id = None
for key, value in [('keywords', keywords),
('key_value_pairs', key_value_pairs),
('data', data)]:
if value:
dct[key] = value
else:
dct.pop(key, None)
if id is None:
id = nextid
ids.append(id)
nextid += 1
bigdct[id] = dct
self._write_json(bigdct, ids, nextid)
return id
def _write(self, atoms, keywords, key_value_pairs, data):
Database._write(self, atoms, keywords, key_value_pairs, data)
bigdct = {}
ids = []
nextid = 1
if isinstance(self.filename, str) and os.path.isfile(self.filename):
try:
bigdct, ids, nextid = self._read_json()
except (SyntaxError, ValueError):
pass
if isinstance(atoms, dict):
dct = dict((key, atoms[key]) for key in reserved_keys
if key in atoms and key != 'id')
unique_id = dct['unique_id']
for id in ids:
if bigdct[id]['unique_id'] == unique_id:
break
else:
id = None
dct['mtime'] = now()
else:
dct = self.collect_data(atoms)
id = None
for key, value in [('keywords', keywords),
('key_value_pairs', key_value_pairs),
('data', data)]:
if value:
dct[key] = value
else:
dct.pop(key, None)
if id is None:
id = nextid
ids.append(id)
nextid += 1
bigdct[id] = dct
self._write_json(bigdct, ids, nextid)
return id
def add_project(db: Database) -> None:
"""Add database to projects with name 'default'."""
all_keys = set()
for row in db.select(columns=['key_value_pairs'], include_data=False):
all_keys.update(row._keys)
kd = {key: (key, '', '') for key in all_keys}
projects['default'] = {
'name': 'default',
'uid_key': 'id',
'key_descriptions': create_key_descriptions(kd),
'database': db,
'row_to_dict_function': row_to_dict,
'handle_query_function': handle_query,
'default_columns': all_columns[:],
'search_template': 'ase/db/templates/search.html',
'row_template': 'ase/db/templates/row.html'}
def create_table(self, db: Database, uid_key: str) -> Table:
query = self.query
if self.nrows is None:
try:
self.nrows = db.count(query)
except (ValueError, KeyError) as e:
error = ', '.join(['Bad query'] + list(e.args))
flash(error)
query = 'id=0' # this will return no rows
self.nrows = 0
table = Table(db, uid_key)
table.select(query, self.columns, self.sort,
self.limit, offset=self.page * self.limit)
table.format()
table.addcolumns = sorted(column for column in
all_columns + table.keys
if column not in table.columns)
return table
def _write(self, atoms, key_value_pairs, data, id):
ext_tables = key_value_pairs.pop("external_tables", {})
Database._write(self, atoms, key_value_pairs, data)
mtime = now()
encode = self.encode
blob = self.blob
if not isinstance(atoms, AtomsRow):
row = AtomsRow(atoms)
row.ctime = mtime
row.user = os.getenv('USER')
else:
row = atoms
# Extract the external tables from AtomsRow
names = self._get_external_table_names()
for name in names:
new_table = row.get(name, {})
if new_table:
ext_tables[name] = new_table
if not id and not key_value_pairs and not ext_tables:
key_value_pairs = row.key_value_pairs
for k, v in ext_tables.items():
dtype = self._guess_type(v)
self._create_table_if_not_exists(k, dtype)
constraints = row._constraints
if constraints:
if isinstance(constraints, list):
constraints = encode(constraints)
else:
constraints = None
values = (row.unique_id, row.ctime, mtime, row.user, blob(row.numbers),
blob(row.positions), blob(row.cell),
int(np.dot(row.pbc,
[1, 2, 4])), blob(row.get('initial_magmoms')),
blob(row.get('initial_charges')), blob(row.get('masses')),
blob(row.get('tags')), blob(row.get('momenta')), constraints)
if 'calculator' in row:
values += (row.calculator, encode(row.calculator_parameters))
else:
values += (None, None)
if not data:
data = row._data
with self.managed_connection() as con:
if not isinstance(data, (str, bytes)):
data = encode(data, binary=self.version >= 9)
values += (row.get('energy'), row.get('free_energy'),
blob(row.get('forces')), blob(row.get('stress')),
blob(row.get('dipole')), blob(row.get('magmoms')),
row.get('magmom'), blob(row.get('charges')),
encode(key_value_pairs), data, len(row.numbers),
float_if_not_none(row.get('fmax')),
float_if_not_none(row.get('smax')),
float_if_not_none(row.get('volume')), float(row.mass),
float(row.charge))
cur = con.cursor()
if id is None:
q = self.default + ', ' + ', '.join('?' * len(values))
cur.execute('INSERT INTO systems VALUES ({})'.format(q),
values)
id = self.get_last_id(cur)
else:
self._delete(cur, [id], [
'keys', 'text_key_values', 'number_key_values', 'species'
])
q = ', '.join(name + '=?' for name in self.columnnames[1:])
cur.execute('UPDATE systems SET {} WHERE id=?'.format(q),
values + (id, ))
count = row.count_atoms()
if count:
species = [(atomic_numbers[symbol], n, id)
for symbol, n in count.items()]
cur.executemany('INSERT INTO species VALUES (?, ?, ?)',
species)
text_key_values = []
number_key_values = []
for key, value in key_value_pairs.items():
if isinstance(value, (numbers.Real, np.bool_)):
number_key_values.append([key, float(value), id])
else:
assert isinstance(value, str)
text_key_values.append([key, value, id])
cur.executemany('INSERT INTO text_key_values VALUES (?, ?, ?)',
text_key_values)
cur.executemany('INSERT INTO number_key_values VALUES (?, ?, ?)',
number_key_values)
cur.executemany('INSERT INTO keys VALUES (?, ?)',
[(key, id) for key in key_value_pairs])
# Insert entries in the valid tables
for tabname in ext_tables.keys():
entries = ext_tables[tabname]
entries['id'] = id
self._insert_in_external_table(cur,
name=tabname,
entries=ext_tables[tabname])
return id
def _write(self, atoms, key_value_pairs, data, id):
Database._write(self, atoms, key_value_pairs, data)
encode = self.encode
con = self.connection or self._connect()
self._initialize(con)
cur = con.cursor()
mtime = now()
blob = self.blob
text_key_values = []
number_key_values = []
if not isinstance(atoms, AtomsRow):
row = AtomsRow(atoms)
row.ctime = mtime
row.user = os.getenv('USER')
else:
row = atoms
if id:
self._delete(cur, [id], ['keys', 'text_key_values',
'number_key_values', 'species'])
else:
if not key_value_pairs:
key_value_pairs = row.key_value_pairs
constraints = row._constraints
if constraints:
if isinstance(constraints, list):
constraints = encode(constraints)
else:
constraints = None
values = (row.unique_id,
row.ctime,
mtime,
row.user,
blob(row.numbers),
blob(row.positions),
blob(row.cell),
int(np.dot(row.pbc, [1, 2, 4])),
blob(row.get('initial_magmoms')),
blob(row.get('initial_charges')),
blob(row.get('masses')),
blob(row.get('tags')),
blob(row.get('momenta')),
constraints)
if 'calculator' in row:
values += (row.calculator, encode(row.calculator_parameters))
else:
values += (None, None)
if not data:
data = row._data
if not isinstance(data, basestring):
data = encode(data)
values += (row.get('energy'),
row.get('free_energy'),
blob(row.get('forces')),
blob(row.get('stress')),
blob(row.get('dipole')),
blob(row.get('magmoms')),
row.get('magmom'),
blob(row.get('charges')),
encode(key_value_pairs),
data,
len(row.numbers),
float_if_not_none(row.get('fmax')),
float_if_not_none(row.get('smax')),
float_if_not_none(row.get('volume')),
float(row.mass),
float(row.charge))
if id is None:
q = self.default + ', ' + ', '.join('?' * len(values))
cur.execute('INSERT INTO systems VALUES ({})'.format(q),
values)
id = self.get_last_id(cur)
else:
q = ', '.join(name + '=?' for name in self.columnnames[1:])
cur.execute('UPDATE systems SET {} WHERE id=?'.format(q),
values + (id,))
count = row.count_atoms()
if count:
species = [(atomic_numbers[symbol], n, id)
for symbol, n in count.items()]
cur.executemany('INSERT INTO species VALUES (?, ?, ?)',
species)
text_key_values = []
number_key_values = []
for key, value in key_value_pairs.items():
if isinstance(value, (numbers.Real, np.bool_)):
number_key_values.append([key, float(value), id])
else:
assert isinstance(value, basestring)
text_key_values.append([key, value, id])
cur.executemany('INSERT INTO text_key_values VALUES (?, ?, ?)',
text_key_values)
cur.executemany('INSERT INTO number_key_values VALUES (?, ?, ?)',
number_key_values)
cur.executemany('INSERT INTO keys VALUES (?, ?)',
[(key, id) for key in key_value_pairs])
if self.connection is None:
con.commit()
con.close()
return id
def _write(self, atoms, keywords, key_value_pairs, data):
Database._write(self, atoms, keywords, key_value_pairs, data)
con = self._connect()
self._initialize(con)
cur = con.cursor()
id = None
if isinstance(atoms, dict):
dct = atoms
unique_id = dct['unique_id']
cur.execute('SELECT id FROM systems WHERE unique_id=?',
(unique_id,))
rows = cur.fetchall()
if rows:
id = rows[0][0]
self._delete(cur, [id], ['keywords', 'text_key_values',
'number_key_values'])
dct['mtime'] = now()
else:
dct = self.collect_data(atoms)
if 'constraints' in dct:
constraints = encode(dct['constraints'])
else:
constraints = None
numbers = dct.get('numbers')
row = (dct['unique_id'],
dct['ctime'],
dct['mtime'],
dct['user'],
blob(numbers),
blob(dct.get('positions')),
blob(dct.get('cell')),
int(np.dot(dct.get('pbc'), [1, 2, 4])),
blob(dct.get('initial_magmoms')),
blob(dct.get('initial_charges')),
blob(dct.get('masses')),
blob(dct.get('tags')),
blob(dct.get('momenta')),
constraints)
if 'calculator' in dct:
row += (dct['calculator'],
encode(dct['calculator_parameters']))
else:
row += (None, None)
magmom = dct.get('magmom')
if magmom is not None:
# magmom can be one or three numbers (non-collinear case)
magmom = np.array(magmom)
row += (dct.get('energy'),
dct.get('free_energy'),
blob(dct.get('forces')),
blob(dct.get('stress')),
blob(dct.get('dipole')),
blob(dct.get('magmoms')),
blob(magmom),
blob(dct.get('charges')),
encode(keywords),
encode(key_value_pairs),
encode(data),
len(numbers))
if id is None:
q = self.default + ', ' + ', '.join('?' * len(row))
cur.execute('INSERT INTO systems VALUES ({0})'.format(q),
row)
else:
q = ', '.join(line.split()[0].lstrip() + '=?'
for line in init_statements[0].splitlines()[2:])
cur.execute('UPDATE systems SET {0} WHERE id=?'.format(q),
row + (id,))
if id is None:
id = self.get_last_id(cur)
if len(numbers) > 0:
count = np.bincount(numbers)
unique_numbers = count.nonzero()[0]
species = [(int(Z), int(count[Z]), id) for Z in unique_numbers]
cur.executemany('INSERT INTO species VALUES (?, ?, ?)',
species)
text_key_values = []
number_key_values = []
for key, value in key_value_pairs.items():
if isinstance(value, (float, int)):
number_key_values.append([key, float(value), id])
else:
assert isinstance(value, (str, unicode))
text_key_values.append([key, value, id])
cur.executemany('INSERT INTO text_key_values VALUES (?, ?, ?)',
text_key_values)
cur.executemany('INSERT INTO number_key_values VALUES (?, ?, ?)',
number_key_values)
cur.executemany('INSERT INTO keywords VALUES (?, ?)',
[(keyword, id) for keyword in keywords])
# Insert keys in keywords table also so that it is easy to query
# for the existance of keys:
cur.executemany('INSERT INTO keywords VALUES (?, ?)',
[(key, id) for key in key_value_pairs])
con.commit()
con.close()
return id
def _write(self, atoms, key_value_pairs, data, id):
ext_tables = key_value_pairs.pop("external_tables", {})
Database._write(self, atoms, key_value_pairs, data)
encode = self.encode
con = self.connection or self._connect()
self._initialize(con)
cur = con.cursor()
mtime = now()
blob = self.blob
text_key_values = []
number_key_values = []
if not isinstance(atoms, AtomsRow):
row = AtomsRow(atoms)
row.ctime = mtime
row.user = os.getenv('USER')
else:
row = atoms
# Extract the external tables from AtomsRow
names = self._get_external_table_names(db_con=con)
for name in names:
new_table = row.get(name, {})
if new_table:
ext_tables[name] = new_table
if id:
self._delete(
cur, [id],
['keys', 'text_key_values', 'number_key_values', 'species'])
else:
if not key_value_pairs:
key_value_pairs = row.key_value_pairs
constraints = row._constraints
if constraints:
if isinstance(constraints, list):
constraints = encode(constraints)
else:
constraints = None
values = (row.unique_id, row.ctime, mtime, row.user, blob(row.numbers),
blob(row.positions), blob(row.cell),
int(np.dot(row.pbc,
[1, 2, 4])), blob(row.get('initial_magmoms')),
blob(row.get('initial_charges')), blob(row.get('masses')),
blob(row.get('tags')), blob(row.get('momenta')), constraints)
if 'calculator' in row:
values += (row.calculator, encode(row.calculator_parameters))
else:
values += (None, None)
if not data:
data = row._data
if not isinstance(data, basestring):
data = encode(data)
values += (row.get('energy'), row.get('free_energy'),
blob(row.get('forces')), blob(row.get('stress')),
blob(row.get('dipole')), blob(row.get('magmoms')),
row.get('magmom'), blob(row.get('charges')),
encode(key_value_pairs), data, len(row.numbers),
float_if_not_none(row.get('fmax')),
float_if_not_none(row.get('smax')),
float_if_not_none(row.get('volume')), float(row.mass),
float(row.charge))
if id is None:
q = self.default + ', ' + ', '.join('?' * len(values))
cur.execute('INSERT INTO systems VALUES ({})'.format(q), values)
id = self.get_last_id(cur)
else:
q = ', '.join(name + '=?' for name in self.columnnames[1:])
cur.execute('UPDATE systems SET {} WHERE id=?'.format(q),
values + (id, ))
count = row.count_atoms()
if count:
species = [(atomic_numbers[symbol], n, id)
for symbol, n in count.items()]
cur.executemany('INSERT INTO species VALUES (?, ?, ?)', species)
text_key_values = []
number_key_values = []
for key, value in key_value_pairs.items():
if isinstance(value, (numbers.Real, np.bool_)):
number_key_values.append([key, float(value), id])
else:
assert isinstance(value, basestring)
text_key_values.append([key, value, id])
cur.executemany('INSERT INTO text_key_values VALUES (?, ?, ?)',
text_key_values)
cur.executemany('INSERT INTO number_key_values VALUES (?, ?, ?)',
number_key_values)
cur.executemany('INSERT INTO keys VALUES (?, ?)',
[(key, id) for key in key_value_pairs])
# Update external tables
valid_entries = []
for k, v in ext_tables.items():
try:
# Guess the type of the value
dtype = self._guess_type(v)
self._create_table_if_not_exists(k, dtype, db_con=con)
v["id"] = id
valid_entries.append(k)
except ValueError as exc:
# Close the connection without committing
if self.connection is None:
con.close()
# Raise error again
raise ValueError(exc)
# Insert entries in the valid tables
for tabname in valid_entries:
try:
self._insert_in_external_table(cur,
name=tabname,
entries=ext_tables[tabname])
except ValueError as exc:
# Close the connection without committing
if self.connection is None:
con.close()
# Raise the error again
raise ValueError(exc)
if self.connection is None:
con.commit()
con.close()
return id
def _write(self, atoms, key_value_pairs, data):
Database._write(self, atoms, key_value_pairs, data)
con = self.connection or self._connect()
self._initialize(con)
cur = con.cursor()
id = None
if not isinstance(atoms, AtomsRow):
row = AtomsRow(atoms)
row.ctime = mtime = now()
row.user = os.getenv('USER')
else:
row = atoms
cur.execute('SELECT id FROM systems WHERE unique_id=?',
(row.unique_id,))
results = cur.fetchall()
if results:
id = results[0][0]
self._delete(cur, [id], ['keys', 'text_key_values',
'number_key_values'])
mtime = now()
constraints = row._constraints
if constraints:
if isinstance(constraints, list):
constraints = encode(constraints)
else:
constraints = None
values = (row.unique_id,
row.ctime,
mtime,
row.user,
blob(row.numbers),
blob(row.positions),
blob(row.cell),
int(np.dot(row.pbc, [1, 2, 4])),
blob(row.get('initial_magmoms')),
blob(row.get('initial_charges')),
blob(row.get('masses')),
blob(row.get('tags')),
blob(row.get('momenta')),
constraints)
if 'calculator' in row:
if not isinstance(row.calculator_parameters, basestring):
row.calculator_parameters = encode(row.calculator_parameters)
values += (row.calculator,
row.calculator_parameters)
else:
values += (None, None)
if key_value_pairs is None:
key_value_pairs = row.key_value_pairs
if not data:
data = row._data
if not isinstance(data, basestring):
data = encode(data)
values += (row.get('energy'),
row.get('free_energy'),
blob(row.get('forces')),
blob(row.get('stress')),
blob(row.get('dipole')),
blob(row.get('magmoms')),
row.get('magmom'),
blob(row.get('charges')),
encode(key_value_pairs),
data,
len(row.numbers),
float_if_not_none(row.get('fmax')),
float_if_not_none(row.get('smax')),
float_if_not_none(row.get('volume')),
float(row.mass),
float(row.charge))
if id is None:
q = self.default + ', ' + ', '.join('?' * len(values))
cur.execute('INSERT INTO systems VALUES ({0})'.format(q),
values)
else:
q = ', '.join(line.split()[0].lstrip() + '=?'
for line in init_statements[0].splitlines()[2:])
cur.execute('UPDATE systems SET {0} WHERE id=?'.format(q),
values + (id,))
if id is None:
id = self.get_last_id(cur)
count = row.count_atoms()
if count:
species = [(atomic_numbers[symbol], n, id)
for symbol, n in count.items()]
cur.executemany('INSERT INTO species VALUES (?, ?, ?)',
species)
text_key_values = []
number_key_values = []
for key, value in key_value_pairs.items():
if isinstance(value, (float, int)):
number_key_values.append([key, float(value), id])
else:
assert isinstance(value, basestring)
text_key_values.append([key, value, id])
cur.executemany('INSERT INTO text_key_values VALUES (?, ?, ?)',
text_key_values)
cur.executemany('INSERT INTO number_key_values VALUES (?, ?, ?)',
number_key_values)
cur.executemany('INSERT INTO keys VALUES (?, ?)',
[(key, id) for key in key_value_pairs])
if self.connection is None:
con.commit()
con.close()
return id