From 3f3ba519ae3c3346945928b21ab36f7238e5387e Mon Sep 17 00:00:00 2001 From: Wojtek Kosior Date: Thu, 11 Aug 2022 09:39:19 +0200 Subject: save computed payloads into sqlite db --- src/hydrilla/item_infos.py | 24 ++--- src/hydrilla/proxy/state_impl/base.py | 38 ++++---- src/hydrilla/proxy/state_impl/concrete_state.py | 113 +++++++++++++++++------- src/hydrilla/proxy/tables.sql | 20 +++-- 4 files changed, 120 insertions(+), 75 deletions(-) (limited to 'src/hydrilla') diff --git a/src/hydrilla/item_infos.py b/src/hydrilla/item_infos.py index c1e1e31..0bdc95e 100644 --- a/src/hydrilla/item_infos.py +++ b/src/hydrilla/item_infos.py @@ -153,14 +153,14 @@ class ItemInfoBase(ABC, ItemIdentity, Categorizable): """....""" type_name: t.ClassVar[str] - source_name: str = dc.field(hash=False) - source_copyright: tuple[FileSpecifier, ...] = dc.field(hash=False) - uuid: t.Optional[str] = dc.field(hash=False) - long_name: str = dc.field(hash=False) - allows_eval: bool = dc.field(hash=False) - allows_cors_bypass: bool = dc.field(hash=False) - required_mappings: tuple[ItemSpecifier, ...] = dc.field(hash=False) - generated_by: t.Optional[GeneratedBy] = dc.field(hash=False) + source_name: str = dc.field(hash=False, compare=False) + source_copyright: tuple[FileSpecifier, ...] = dc.field(hash=False, compare=False) + uuid: t.Optional[str] = dc.field(hash=False, compare=False) + long_name: str = dc.field(hash=False, compare=False) + allows_eval: bool = dc.field(hash=False, compare=False) + allows_cors_bypass: bool = dc.field(hash=False, compare=False) + required_mappings: tuple[ItemSpecifier, ...] = dc.field(hash=False, compare=False) + generated_by: t.Optional[GeneratedBy] = dc.field(hash=False, compare=False) # def path_relative_to_type(self) -> str: # """ @@ -237,9 +237,9 @@ class ResourceInfo(ItemInfoBase): """....""" type_name: t.ClassVar[str] = 'resource' - revision: int = dc.field(hash=False) - dependencies: tuple[ItemSpecifier, ...] = dc.field(hash=False) - scripts: tuple[FileSpecifier, ...] = dc.field(hash=False) + revision: int = dc.field(hash=False, compare=False) + dependencies: tuple[ItemSpecifier, ...] = dc.field(hash=False, compare=False) + scripts: tuple[FileSpecifier, ...] = dc.field(hash=False, compare=False) @property def versioned_identifier(self): @@ -321,7 +321,7 @@ class MappingInfo(ItemInfoBase): """....""" type_name: t.ClassVar[str] = 'mapping' - payloads: t.Mapping[ParsedPattern, ItemSpecifier] = dc.field(hash=False) + payloads: t.Mapping[ParsedPattern, ItemSpecifier] = dc.field(hash=False, compare=False) @staticmethod def make( diff --git a/src/hydrilla/proxy/state_impl/base.py b/src/hydrilla/proxy/state_impl/base.py index 78a50c0..788a93d 100644 --- a/src/hydrilla/proxy/state_impl/base.py +++ b/src/hydrilla/proxy/state_impl/base.py @@ -72,40 +72,38 @@ class HaketiloStateWithFields(state.HaketiloState): lock: threading.RLock = dc.field(default_factory=threading.RLock) @contextmanager - def cursor(self, lock: bool = False, transaction: bool = False) \ + def cursor(self, transaction: bool = False) \ -> t.Iterator[sqlite3.Cursor]: """....""" start_transaction = transaction and not self.connection.in_transaction - assert lock or not start_transaction - try: - if lock: - self.lock.acquire() + self.lock.acquire() if self.current_cursor is not None: yield self.current_cursor return - self.current_cursor = self.connection.cursor() - - if start_transaction: - self.current_cursor.execute('BEGIN TRANSACTION;') - try: - yield self.current_cursor + self.current_cursor = self.connection.cursor() if start_transaction: - self.current_cursor.execute('COMMIT TRANSACTION;') - except: - if start_transaction: - self.current_cursor.execute('ROLLBACK TRANSACTION;') - raise + self.current_cursor.execute('BEGIN TRANSACTION;') + + try: + yield self.current_cursor + + if start_transaction: + assert self.connection.in_transaction + self.current_cursor.execute('COMMIT TRANSACTION;') + except: + if start_transaction: + self.current_cursor.execute('ROLLBACK TRANSACTION;') + raise + finally: + self.current_cursor = None finally: - self.current_cursor = None - - if lock: - self.lock.release() + self.lock.release() def rebuild_structures(self, cursor: sqlite3.Cursor) -> None: """....""" diff --git a/src/hydrilla/proxy/state_impl/concrete_state.py b/src/hydrilla/proxy/state_impl/concrete_state.py index 1b46ae9..ccb1269 100644 --- a/src/hydrilla/proxy/state_impl/concrete_state.py +++ b/src/hydrilla/proxy/state_impl/concrete_state.py @@ -313,19 +313,10 @@ def get_or_make_item_version( cursor: sqlite3.Cursor, item_id: int, repo_iteration_id: int, - definition: str, - info: item_infos.AnyInfo + version: versions.VerTuple, + definition: str ) -> int: - ver_str = versions.version_string(info.version) - - values = ( - item_id, - ver_str, - repo_iteration_id, - definition, - info.allows_eval, - info.allows_cors_bypass - ) + ver_str = versions.version_string(version) cursor.execute( ''' @@ -333,13 +324,11 @@ def get_or_make_item_version( item_id, version, repo_iteration_id, - definition, - eval_allowed, - cors_bypass_allowed + definition ) - VALUES(?, ?, ?, ?, ?, ?); + VALUES(?, ?, ?, ?); ''', - values + (item_id, ver_str, repo_iteration_id, definition) ) cursor.execute( @@ -417,12 +406,12 @@ def make_file_use( (item_version_id, file_id, name, type, mime_type, idx) ) -def get_infos_of_type(cursor: sqlite3.Cursor, info_type: t.Type[AnyInfoVar]) \ - -> t.Iterable[AnyInfoVar]: +def get_infos_of_type(cursor: sqlite3.Cursor, info_type: t.Type[AnyInfoVar],) \ + -> t.Mapping[AnyInfoVar, int]: cursor.execute( ''' SELECT - iv.definition, r.name, ri.iteration + i.item_id, iv.definition, r.name, ri.iteration FROM item_versions AS iv JOIN items AS i USING (item_id) @@ -434,11 +423,12 @@ def get_infos_of_type(cursor: sqlite3.Cursor, info_type: t.Type[AnyInfoVar]) \ (info_type.type_name[0].upper(),) ) - result: list[AnyInfoVar] = [] + result: dict[AnyInfoVar, int] = {} - for definition, repo_name, repo_iteration in cursor.fetchall(): + for item_id, definition, repo_name, repo_iteration in cursor.fetchall(): definition_io = io.StringIO(definition) - result.append(info_type.load(definition_io, repo_name, repo_iteration)) + info = info_type.load(definition_io, repo_name, repo_iteration) + result[info] = item_id return result @@ -449,7 +439,7 @@ class ConcreteHaketiloState(base.HaketiloStateWithFields): self._populate_database_with_stuff_from_temporary_malcontent_dir() - with self.cursor() as cursor: + with self.cursor(transaction=True) as cursor: self.rebuild_structures(cursor) def _prepare_database(self) -> None: @@ -459,9 +449,12 @@ class ConcreteHaketiloState(base.HaketiloStateWithFields): try: cursor.execute( ''' - SELECT COUNT(name) - FROM sqlite_master - WHERE name = 'general' AND type = 'table'; + SELECT + COUNT(name) + FROM + sqlite_master + WHERE + name = 'general' AND type = 'table'; ''' ) @@ -473,8 +466,10 @@ class ConcreteHaketiloState(base.HaketiloStateWithFields): else: cursor.execute( ''' - SELECT haketilo_version - FROM general; + SELECT + haketilo_version + FROM + general; ''' ) @@ -495,7 +490,7 @@ class ConcreteHaketiloState(base.HaketiloStateWithFields): malcontent_dir_path = self.store_dir / 'temporary_malcontent' files_by_sha256_path = malcontent_dir_path / 'file' / 'sha256' - with self.cursor(lock=True, transaction=True) as cursor: + with self.cursor(transaction=True) as cursor: for info_type in [item_infos.ResourceInfo, item_infos.MappingInfo]: info: item_infos.AnyInfo for info, definition in read_items( @@ -517,8 +512,8 @@ class ConcreteHaketiloState(base.HaketiloStateWithFields): cursor, item_id, repo_iteration_id, - definition, - info + info.version, + definition ) if info_type is item_infos.MappingInfo: @@ -574,16 +569,66 @@ class ConcreteHaketiloState(base.HaketiloStateWithFields): ) def rebuild_structures(self, cursor: sqlite3.Cursor) -> None: + assert self.connection.in_transaction + resources = get_infos_of_type(cursor, item_infos.ResourceInfo) mappings = get_infos_of_type(cursor, item_infos.MappingInfo) - payloads = compute_payloads(resources, mappings) + payloads = compute_payloads(resources.keys(), mappings.keys()) payloads_data = {} + cursor.execute('DELETE FROM payloads;') + for mapping_info, by_pattern in payloads.items(): for num, (pattern, payload) in enumerate(by_pattern.items()): - payload_id = f'{num}@{mapping_info.identifier}' + print('adding payload') + cursor.execute( + ''' + INSERT INTO payloads( + mapping_item_id, + pattern, + eval_allowed, + cors_bypass_allowed + ) + VALUES (?, ?, ?, ?); + ''', + ( + mappings[mapping_info], + pattern.orig_url, + payload.allows_eval, + payload.allows_cors_bypass + ) + ) + + cursor.execute( + ''' + SELECT + payload_id + FROM + payloads + WHERE + mapping_item_id = ? AND pattern = ?; + ''', + (mappings[mapping_info], pattern.orig_url) + ) + + (payload_id_int,), = cursor.fetchall() + + for res_num, resource_info in enumerate(payload.resources): + cursor.execute( + ''' + INSERT INTO resolved_depended_resources( + payload_id, + resource_item_id, + idx + ) + VALUES(?, ?, ?); + ''', + (payload_id_int, resources[resource_info], res_num) + ) + + payload_id = str(payload_id_int) ref = ConcretePayloadRef(payload_id, payload) diff --git a/src/hydrilla/proxy/tables.sql b/src/hydrilla/proxy/tables.sql index 53539a7..25493d3 100644 --- a/src/hydrilla/proxy/tables.sql +++ b/src/hydrilla/proxy/tables.sql @@ -141,10 +141,6 @@ CREATE TABLE item_versions( version VARCHAR NOT NULL, repo_iteration_id INTEGER NOT NULL, definition TEXT NOT NULL, - -- What privileges should be granted on pages where this - -- resource/mapping is used. - eval_allowed BOOLEAN NOT NULL, - cors_bypass_allowed BOOLEAN NOT NULL, UNIQUE (item_id, version, repo_iteration_id), -- Allow foreign key from "mapping_statuses". @@ -157,15 +153,19 @@ CREATE TABLE item_versions( ); CREATE TABLE payloads( - payload_id INTEGER PRIMARY KEY, + payload_id INTEGER PRIMARY KEY, - mapping_item_id INTEGER NOT NULL, - pattern VARCHAR NOT NULL, + mapping_item_id INTEGER NOT NULL, + pattern VARCHAR NOT NULL, + -- What privileges should be granted on pages where this + -- resource/mapping is used. + eval_allowed BOOLEAN NOT NULL, + cors_bypass_allowed BOOLEAN NOT NULL, UNIQUE (mapping_item_id, pattern), FOREIGN KEY (mapping_item_id) - REFERENCES item_versions (versioned_item_id) + REFERENCES item_versions (item_version_id) ON DELETE CASCADE ); @@ -179,9 +179,11 @@ CREATE TABLE resolved_depended_resources( PRIMARY KEY (payload_id, resource_item_id), FOREIGN KEY (payload_id) - REFERENCES payloads (payload_id), + REFERENCES payloads (payload_id) + ON DELETE CASCADE, FOREIGN KEY (resource_item_id) REFERENCES item_versions (item_version_id) + ON DELETE CASCADE ) WITHOUT ROWID; -- CREATE TABLE resolved_required_mappings( -- cgit v1.2.3