linkml-store 0.1.8__py3-none-any.whl → 0.1.9__py3-none-any.whl

linkml_store/api/client.py

@@ -9,6 +9,7 @@ from linkml_store.api import Database
 from linkml_store.api.config import ClientConfig
 from linkml_store.api.stores.chromadb.chromadb_database import ChromaDBDatabase
 from linkml_store.api.stores.duckdb.duckdb_database import DuckDBDatabase
+from linkml_store.api.stores.filesystem.filesystem_database import FileSystemDatabase
 from linkml_store.api.stores.mongodb.mongodb_database import MongoDBDatabase
 from linkml_store.api.stores.solr.solr_database import SolrDatabase
 
@@ -20,6 +21,7 @@ HANDLE_MAP = {
     "solr": SolrDatabase,
     "mongodb": MongoDBDatabase,
     "chromadb": ChromaDBDatabase,
+    "file": FileSystemDatabase,
 }
 
 

linkml_store/api/collection.py

@@ -4,16 +4,19 @@ import hashlib
 import logging
 from collections import defaultdict
 from pathlib import Path
-from typing import TYPE_CHECKING, Any, Dict, Iterator, List, Optional, TextIO, Type, Union
+from typing import TYPE_CHECKING, Any, Dict, Generic, Iterator, List, Optional, TextIO, Tuple, Type, Union
 
 import numpy as np
+from linkml_runtime import SchemaView
 from linkml_runtime.linkml_model import ClassDefinition, SlotDefinition
 from linkml_runtime.linkml_model.meta import ArrayExpression
 from pydantic import BaseModel
 
+from linkml_store.api.types import DatabaseType
 from linkml_store.index import get_indexer
 from linkml_store.utils.format_utils import load_objects
 from linkml_store.utils.object_utils import clean_empties
+from linkml_store.utils.patch_utils import PatchDict, apply_patches_to_list, patches_from_objects_lists
 
 try:
     from linkml.validator.report import ValidationResult
@@ -36,7 +39,7 @@ IDENTIFIER = str
 FIELD_NAME = str
 
 
-class Collection:
+class Collection(Generic[DatabaseType]):
     """
     A collection is an organized set of objects of the same or similar type.
 
@@ -56,7 +59,7 @@ class Collection:
     """
 
     # name: str
-    parent: Optional["Database"] = None
+    parent: Optional[DatabaseType] = None
     _indexers: Optional[Dict[str, Indexer]] = None
     # hidden: Optional[bool] = False
 
@@ -197,6 +200,10 @@ class Collection:
         """
         raise NotImplementedError
 
+    def _post_insert_hook(self, objs: List[OBJECT], **kwargs):
+        patches = [{"op": "add", "path": "/0", "value": obj} for obj in objs]
+        self._broadcast(patches, **kwargs)
+
     def delete(self, objs: Union[OBJECT, List[OBJECT]], **kwargs) -> Optional[int]:
         """
         Delete one or more objects from the collection.
@@ -301,7 +308,7 @@ class Collection:
 
     def query_facets(
         self, where: Optional[Dict] = None, facet_columns: List[str] = None, facet_limit=DEFAULT_FACET_LIMIT, **kwargs
-    ) -> Dict[str, Dict[str, int]]:
+    ) -> Dict[str, List[Tuple[Any, int]]]:
         """
         Run a query to get facet counts for one or more columns.
 
@@ -319,7 +326,7 @@ class Collection:
         :param query: A Query object representing the base query.
         :param facet_columns: A list of column names to get facet counts for.
         :param facet_limit:
-        :return: A dictionary where keys are column names and values are pandas DataFrames
+        :return: A dictionary where keys are column names and values are tuples
                  containing the facet counts for each unique value in the respective column.
         """
         raise NotImplementedError
@@ -523,6 +530,7 @@ class Collection:
                 ix_coll.delete_where()
 
         ix_coll.insert(objects_with_ix, **kwargs)
+        ix_coll.commit()
 
     def list_index_names(self) -> List[str]:
         """
@@ -557,12 +565,22 @@ class Collection:
 
         :return:
         """
-        sv = self.parent.schema_view
+        sv: SchemaView = self.parent.schema_view
         if sv:
             cls = sv.get_class(self.target_class_name)
+            if cls and not cls.attributes:
+                if not sv.class_induced_slots(cls.name):
+                    for att in self._induce_attributes():
+                        cls.attributes[att.name] = att
+                    sv.set_modified()
             return cls
         return None
 
+    def _induce_attributes(self) -> List[SlotDefinition]:
+        result = self.find({}, limit=-1)
+        cd = self.induce_class_definition_from_objects(result.rows, max_sample_size=None)
+        return list(cd.attributes.values())
+
     @property
     def identifier_attribute_name(self) -> Optional[str]:
         """
@@ -579,6 +597,37 @@ class Collection:
                     return att.name
         return None
 
+    def set_identifier_attribute_name(self, name: str):
+        """
+        Set the name of the identifier attribute for the collection.
+
+        AKA the primary key.
+
+        :param name: The name of the identifier attribute.
+        """
+        cd = self.class_definition()
+        if not cd:
+            raise ValueError(f"Cannot find class definition for {self.target_class_name}")
+        id_att = None
+        candidates = []
+        sv: SchemaView = self.parent.schema_view
+        cls = sv.get_class(cd.name)
+        existing_id_slot = sv.get_identifier_slot(cls.name)
+        if existing_id_slot:
+            if existing_id_slot.name == name:
+                return
+            existing_id_slot.identifier = False
+        for att in cls.attributes.values():
+            candidates.append(att.name)
+            if att.name == name:
+                att.identifier = True
+                id_att = att
+            else:
+                att.identifier = False
+        if not id_att:
+            raise ValueError(f"No attribute found with name {name} in {candidates}")
+        sv.set_modified()
+
     def object_identifier(self, obj: OBJECT, auto=True) -> Optional[IDENTIFIER]:
         """
         Return the identifier for an object.
@@ -622,6 +671,8 @@ class Collection:
             for k, v in obj.items():
                 keys[k].append(v)
         for k, vs in keys.items():
+            if k == "_id":
+                continue
             multivalueds = []
             inlineds = []
             rngs = []
@@ -698,6 +749,39 @@ class Collection:
         """
         raise NotImplementedError
 
+    def apply_patches(self, patches: List[PatchDict], **kwargs):
+        """
+        Apply a patch to the collection.
+
+        Patches conform to the JSON Patch format,
+
+        :param patches:
+        :param kwargs:
+        :return:
+        """
+        all_objs = self.find(limit=-1).rows
+        primary_key = self.identifier_attribute_name
+        if not primary_key:
+            raise ValueError(f"No primary key for {self.target_class_name}")
+        new_objs = apply_patches_to_list(all_objs, patches, primary_key=primary_key, **kwargs)
+        self.replace(new_objs)
+
+    def diff(self, other: "Collection", **kwargs):
+        """
+        Diff two collections.
+
+        :param other:
+        :param kwargs:
+        :return:
+        """
+        src_objs = self.find(limit=-1).rows
+        tgt_objs = other.find(limit=-1).rows
+        primary_key = self.identifier_attribute_name
+        if not primary_key:
+            raise ValueError(f"No primary key for {self.target_class_name}")
+        patches_from_objects_lists(src_objs, tgt_objs, primary_key=primary_key)
+        return patches_from_objects_lists(src_objs, tgt_objs, primary_key=primary_key)
+
     def iter_validate_collection(self, **kwargs) -> Iterator["ValidationResult"]:
         """
         Validate the contents of the collection
@@ -717,3 +801,14 @@ class Collection:
         for obj in result.rows:
             obj = clean_empties(obj)
             yield from validator.iter_results(obj, class_name)
+
+    def commit(self):
+        """
+        Commit changes to the collection.
+
+        :return:
+        """
+        pass
+
+    def _broadcast(self, *args, **kwargs):
+        self.parent.broadcast(self, *args, **kwargs)

linkml_store/api/database.py

@@ -3,9 +3,24 @@ from abc import ABC
 from collections import defaultdict
 from copy import copy
 from pathlib import Path
-from typing import TYPE_CHECKING, Any, ClassVar, Dict, Iterator, Optional, Sequence, Type, Union
-
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Callable,
+    ClassVar,
+    Dict,
+    Generic,
+    Iterator,
+    List,
+    Optional,
+    Sequence,
+    Type,
+    Union,
+)
+
+from linkml_store.api.types import CollectionType
 from linkml_store.utils.format_utils import load_objects, render_output
+from linkml_store.utils.patch_utils import PatchDict
 
 try:
     from linkml.validator.report import Severity, ValidationResult
@@ -24,8 +39,10 @@ if TYPE_CHECKING:
 
 logger = logging.getLogger(__name__)
 
+LISTENER = Callable[[Collection, List[PatchDict]], None]
+
 
-class Database(ABC):
+class Database(ABC, Generic[CollectionType]):
     """
     A Database provides access to named collections of data.
 
@@ -89,6 +106,8 @@ class Database(ABC):
     metadata: Optional[DatabaseConfig] = None
     collection_class: ClassVar[Optional[Type[Collection]]] = None
 
+    listeners: Optional[List[LISTENER]] = None
+
     def __init__(self, handle: Optional[str] = None, metadata: Optional[DatabaseConfig] = None, **kwargs):
         if metadata:
             self.metadata = metadata
@@ -233,7 +252,8 @@ class Database(ABC):
         :param kwargs:
         :return:
         """
-        raise NotImplementedError()
+        for coll in self.list_collections():
+            coll.commit()
 
     def close(self, **kwargs):
         """
@@ -301,6 +321,7 @@ class Database(ABC):
             alias = name
         self._collections[alias] = collection
         if recreate_if_exists:
+            logger.debug(f"Recreating collection {collection.name}")
             collection.delete_where({}, missing_ok=True)
         return collection
 
@@ -418,7 +439,11 @@ class Database(ABC):
         :return:
 
         """
-        raise NotImplementedError
+        if query.from_table:
+            collection = self.get_collection(query.from_table)
+            return collection.query(query, **kwargs)
+        else:
+            raise NotImplementedError(f"Querying without a table is not supported in {self.__class__.__name__}")
 
     @property
     def schema_view(self) -> SchemaView:
@@ -689,3 +714,9 @@ class Database(ABC):
         logger.info(f"Exporting object with {len(obj)} collections to {location} in {target_format} format")
         with open(location, "w", encoding="utf-8") as stream:
             stream.write(render_output(obj, format=target_format))
+
+    def broadcast(self, source: Collection, patches: List[PatchDict]):
+        if not self.listeners:
+            return
+        for listener in self.listeners:
+            listener(source, patches)

linkml_store/api/stores/duckdb/duckdb_collection.py

@@ -38,6 +38,7 @@ class DuckDBCollection(Collection):
             with conn.begin():
                 conn.execute(insert(table), objs)
             conn.commit()
+        self._post_insert_hook(objs)
 
     def delete(self, objs: Union[OBJECT, List[OBJECT]], **kwargs) -> Optional[int]:
         if not isinstance(objs, list):

linkml_store/api/stores/filesystem/__init__.py

@@ -1,16 +1,15 @@
 """
-Adapter for DuckDB embedded database.
+Adapter for FileSystem wrapper
 
 Handles have the form:
 
- - ``duckdb:///<path>`` for a file-based database
- - ``duckdb:///:memory:`` for an in-memory database
-"""
+ - ``file:<path>`` for a local file
+ """
 
-from linkml_store.api.stores.duckdb.duckdb_collection import DuckDBCollection
-from linkml_store.api.stores.duckdb.duckdb_database import DuckDBDatabase
+from linkml_store.api.stores.filesystem.filesystem_collection import FileSystemCollection
+from linkml_store.api.stores.filesystem.filesystem_database import FileSystemDatabase
 
 __all__ = [
-    "DuckDBCollection",
-    "DuckDBDatabase",
+    "FileSystemCollection",
+    "FileSystemDatabase",
 ]

linkml_store/api/stores/filesystem/filesystem_collection.py

@@ -1,142 +1,177 @@
 import logging
+from pathlib import Path
 from typing import Any, Dict, List, Optional, Union
 
-import sqlalchemy as sqla
-from linkml_runtime.linkml_model import ClassDefinition, SlotDefinition
-from sqlalchemy import Column, Table, delete, insert, inspect, text
-from sqlalchemy.sql.ddl import CreateTable
-
 from linkml_store.api import Collection
 from linkml_store.api.collection import DEFAULT_FACET_LIMIT, OBJECT
-from linkml_store.api.queries import Query
-from linkml_store.api.stores.duckdb.mappings import TMAP
-from linkml_store.utils.sql_utils import facet_count_sql
+from linkml_store.api.queries import Query, QueryResult
+from linkml_store.api.types import DatabaseType
+from linkml_store.utils.query_utils import mongo_query_to_match_function
 
 logger = logging.getLogger(__name__)
 
 
-class FileSystemCollection(Collection):
-    _table_created: bool = None
+class FileSystemCollection(Collection[DatabaseType]):
+    path: Optional[Path] = None
+    file_format: Optional[str] = None
+    encoding: Optional[str] = None
+    _objects_list: List[OBJECT] = None
+    _object_map: Dict[str, OBJECT] = None
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+        parent: DatabaseType = self.parent
+        if not self.path:
+            if self.parent:
+                self.path = Path(parent.directory_path)
+        self._objects_list = []
+        self._object_map = {}
+        if not self.file_format:
+            self.file_format = "json"
+
+    @property
+    def path_to_file(self):
+        return Path(self.parent.directory_path) / f"{self.name}.{self.file_format}"
+
+    @property
+    def objects_as_list(self) -> List[OBJECT]:
+        if self._object_map:
+            return list(self._object_map.values())
+        else:
+            return self._objects_list
+
+    def _set_objects(self, objs: List[OBJECT]):
+        pk = self.identifier_attribute_name
+        if pk:
+            self._object_map = {obj[pk]: obj for obj in objs}
+            self._objects_list = []
+        else:
+            self._objects_list = objs
+            self._object_map = {}
+
+    def commit(self):
+        path = self.path_to_file
+        if not path:
+            raise ValueError("Path not set")
+        path.parent.mkdir(parents=True, exist_ok=True)
+        self._save(path)
+
+    def _save(self, path: Path):
+        encoding = self.encoding or "utf-8"
+        fmt = self.file_format or "json"
+        mode = "w"
+        if fmt == "parquet":
+            mode = "wb"
+            encoding = None
+        with open(path, mode, encoding=encoding) as stream:
+            if fmt == "json":
+                import json
+
+                json.dump(self.objects_as_list, stream, indent=2)
+            elif fmt == "jsonl":
+                import jsonlines
+
+                writer = jsonlines.Writer(stream)
+                writer.write_all(self.objects_as_list)
+            elif fmt == "yaml":
+                import yaml
+
+                yaml.dump_all(self.objects_as_list, stream)
+            elif fmt == "parquet":
+                import pandas as pd
+                import pyarrow
+                import pyarrow.parquet as pq
+
+                df = pd.DataFrame(self.objects_as_list)
+                table = pyarrow.Table.from_pandas(df)
+                pq.write_table(table, stream)
+            elif fmt in {"csv", "tsv"}:
+                import csv
+
+                delimiter = "\t" if fmt == "tsv" else ","
+                fieldnames = list(self.objects_as_list[0].keys())
+                for obj in self.objects_as_list[1:]:
+                    fieldnames.extend([k for k in obj.keys() if k not in fieldnames])
+                writer = csv.DictWriter(stream, fieldnames=fieldnames, delimiter=delimiter)
+                writer.writeheader()
+                for obj in self.objects_as_list:
+                    writer.writerow(obj)
+            else:
+                raise ValueError(f"Unsupported file format: {fmt}")
 
     def insert(self, objs: Union[OBJECT, List[OBJECT]], **kwargs):
         if not isinstance(objs, list):
             objs = [objs]
         if not objs:
             return
-        cd = self.class_definition()
-        if not cd:
-            cd = self.induce_class_definition_from_objects(objs)
-        self._create_table(cd)
-        table = self._sqla_table(cd)
-        logger.info(f"Inserting into: {self.alias} // T={table.name}")
-        engine = self.parent.engine
-        col_names = [c.name for c in table.columns]
-        objs = [{k: obj.get(k, None) for k in col_names} for obj in objs]
-        with engine.connect() as conn:
-            with conn.begin():
-                conn.execute(insert(table), objs)
-            conn.commit()
+        pk = self.identifier_attribute_name
+        if pk:
+            for obj in objs:
+                if pk not in obj:
+                    raise ValueError(f"Primary key {pk} not found in object {obj}")
+                pk_val = obj[pk]
+                self._object_map[pk_val] = obj
+        else:
+            self._objects_list.extend(objs)
 
     def delete(self, objs: Union[OBJECT, List[OBJECT]], **kwargs) -> Optional[int]:
         if not isinstance(objs, list):
             objs = [objs]
-        cd = self.class_definition()
-        if not cd:
-            cd = self.induce_class_definition_from_objects(objs)
-        table = self._sqla_table(cd)
-        engine = self.parent.engine
-        with engine.connect() as conn:
+        if not objs:
+            return 0
+        pk = self.identifier_attribute_name
+        n = 0
+        if pk:
             for obj in objs:
-                conditions = [table.c[k] == v for k, v in obj.items() if k in cd.attributes]
-                stmt = delete(table).where(*conditions)
-                stmt = stmt.compile(engine)
-                conn.execute(stmt)
-                conn.commit()
-        return
+                pk_val = obj[pk]
+                if pk_val in self._object_map:
+                    del self._object_map[pk_val]
+                    n += 1
+        else:
+            n = len(objs)
+            self._objects_list = [o for o in self._objects_list if o not in objs]
+            n = n - len(objs)
+        return n
 
     def delete_where(self, where: Optional[Dict[str, Any]] = None, missing_ok=True, **kwargs) -> Optional[int]:
         logger.info(f"Deleting from {self.target_class_name} where: {where}")
         if where is None:
             where = {}
-        cd = self.class_definition()
-        if not cd:
-            logger.info(f"No class definition found for {self.target_class_name}, assuming not prepopulated")
-            return 0
-        table = self._sqla_table(cd)
-        engine = self.parent.engine
-        inspector = inspect(engine)
-        table_exists = table.name in inspector.get_table_names()
-        if not table_exists:
-            logger.info(f"Table {table.name} does not exist, assuming no data")
-            return 0
-        with engine.connect() as conn:
-            conditions = [table.c[k] == v for k, v in where.items()]
-            stmt = delete(table).where(*conditions)
-            stmt = stmt.compile(engine)
-            result = conn.execute(stmt)
-            deleted_rows_count = result.rowcount
-            if deleted_rows_count == 0 and not missing_ok:
-                raise ValueError(f"No rows found for {where}")
-            conn.commit()
-            return deleted_rows_count if deleted_rows_count > -1 else None
+
+        def matches(obj: OBJECT):
+            for k, v in where.items():
+                if obj.get(k) != v:
+                    return False
+            return True
+
+        print(type(self))
+        print(self)
+        print(vars(self))
+        curr_objects = [o for o in self.objects_as_list if not matches(o)]
+        self._set_objects(curr_objects)
+
+    def query(self, query: Query, **kwargs) -> QueryResult:
+
+        where = query.where_clause or {}
+        match = mongo_query_to_match_function(where)
+        rows = [o for o in self.objects_as_list if match(o)]
+        count = len(rows)
+        return QueryResult(query=query, num_rows=count, rows=rows)
 
     def query_facets(
         self, where: Dict = None, facet_columns: List[str] = None, facet_limit=DEFAULT_FACET_LIMIT, **kwargs
     ) -> Dict[str, Dict[str, int]]:
-        results = {}
-        cd = self.class_definition()
-        with self.parent.engine.connect() as conn:
-            if not facet_columns:
-                facet_columns = list(self.class_definition().attributes.keys())
-            for col in facet_columns:
-                logger.debug(f"Faceting on {col}")
-                if isinstance(col, tuple):
-                    sd = SlotDefinition(name="PLACEHOLDER")
-                else:
-                    sd = cd.attributes[col]
-                facet_query = self._create_query(where_clause=where)
-                facet_query_str = facet_count_sql(facet_query, col, multivalued=sd.multivalued)
-                logger.debug(f"Facet query: {facet_query_str}")
-                rows = list(conn.execute(text(facet_query_str)))
-                results[col] = rows
-            return results
-
-    def _sqla_table(self, cd: ClassDefinition) -> Table:
-        schema_view = self.parent.schema_view
-        metadata_obj = sqla.MetaData()
-        cols = []
-        for att in schema_view.class_induced_slots(cd.name):
-            typ = TMAP.get(att.range, sqla.String)
-            if att.inlined:
-                typ = sqla.JSON
-            if att.multivalued:
-                typ = sqla.ARRAY(typ, dimensions=1)
-            if att.array:
-                typ = sqla.ARRAY(typ, dimensions=1)
-            col = Column(att.name, typ)
-            cols.append(col)
-        t = Table(self.alias, metadata_obj, *cols)
-        return t
-
-    def _create_table(self, cd: ClassDefinition):
-        if self._table_created or self.metadata.is_prepopulated:
-            logger.info(f"Already have table for: {cd.name}")
-            return
-        query = Query(
-            from_table="information_schema.tables", where_clause={"table_type": "BASE TABLE", "table_name": self.alias}
-        )
-        qr = self.parent.query(query)
-        if qr.num_rows > 0:
-            logger.info(f"Table already exists for {cd.name}")
-            self._table_created = True
-            self.metadata.is_prepopulated = True
-            return
-        logger.info(f"Creating table for {cd.name}")
-        t = self._sqla_table(cd)
-        ct = CreateTable(t)
-        ddl = str(ct.compile(self.parent.engine))
-        with self.parent.engine.connect() as conn:
-            conn.execute(text(ddl))
-            conn.commit()
-        self._table_created = True
-        self.metadata.is_prepopulated = True
+        match = mongo_query_to_match_function(where)
+        rows = [o for o in self.objects_as_list if match(o)]
+        if not facet_columns:
+            facet_columns = self.class_definition().attributes.keys()
+        facet_results = {c: {} for c in facet_columns}
+        for row in rows:
+            for fc in facet_columns:
+                if fc in row:
+                    v = row[fc]
+                    if v not in facet_results[fc]:
+                        facet_results[fc][v] = 1
+                    else:
+                        facet_results[fc][v] += 1
+        return {fc: list(facet_results[fc].items()) for fc in facet_results}