kaqing 2.0.110__py3-none-any.whl → 2.0.184__py3-none-any.whl

adam/commands/export/drop_export_databases.py

@@ -0,0 +1,37 @@
+from adam.commands.command import Command
+from adam.commands.export.export_databases import export_db
+from adam.repl_state import ReplState
+
+class DropExportDatabases(Command):
+    COMMAND = 'drop all export databases'
+
+    # the singleton pattern
+    def __new__(cls, *args, **kwargs):
+        if not hasattr(cls, 'instance'): cls.instance = super(DropExportDatabases, cls).__new__(cls)
+
+        return cls.instance
+
+    def __init__(self, successor: Command=None):
+        super().__init__(successor)
+
+    def command(self):
+        return DropExportDatabases.COMMAND
+
+    def required(self):
+        return [ReplState.C, ReplState.X]
+
+    def run(self, cmd: str, state: ReplState):
+        if not(args := self.args(cmd)):
+            return super().run(cmd, state)
+
+        with self.validate(args, state) as (args, state):
+            with export_db(state) as dbs:
+                dbs.drop_all()
+
+            return state
+
+    def completion(self, _: ReplState):
+        return {}
+
+    def help(self, _: ReplState):
+        return f'{DropExportDatabases.COMMAND}\t drop all export databases'

adam/commands/export/export.py

@@ -0,0 +1,53 @@
+from adam.commands import extract_options
+from adam.commands.command import Command
+from adam.commands.cql.utils_cql import cassandra_keyspaces, cassandra_table_names
+from adam.commands.export.export_databases import ExportDatabases
+from adam.commands.export.exporter import export
+from adam.repl_state import ReplState, RequiredState
+from adam.sql.sql_completer import SqlCompleter, SqlVariant
+from adam.utils import log
+from adam.utils_k8s.statefulsets import StatefulSets
+
+class ExportTables(Command):
+    COMMAND = 'export'
+
+    # the singleton pattern
+    def __new__(cls, *args, **kwargs):
+        if not hasattr(cls, 'instance'): cls.instance = super(ExportTables, cls).__new__(cls)
+
+        return cls.instance
+
+    def __init__(self, successor: Command=None):
+        super().__init__(successor)
+
+    def command(self):
+        return ExportTables.COMMAND
+
+    def required(self):
+        return RequiredState.CLUSTER_OR_POD
+
+    def run(self, cmd: str, state: ReplState):
+        if not(args := self.args(cmd)):
+            return super().run(cmd, state)
+
+        with self.validate(args, state) as (args, state):
+            with extract_options(args, '--export-only') as (args, export_only):
+                with export(state) as exporter:
+                    return exporter.export(args, export_only=export_only)
+
+    def completion(self, state: ReplState):
+        def sc():
+            return SqlCompleter(lambda: cassandra_table_names(state), expandables={
+                'dml':'export',
+                'columns': lambda table: ['id', '*'],
+                'keyspaces': lambda: cassandra_keyspaces(state),
+                'export-dbs': lambda: ExportDatabases.database_names(),
+            }, variant=SqlVariant.CQL)
+
+        if super().completion(state):
+            return {f'@{p}': {ExportTables.COMMAND: sc()} for p in StatefulSets.pod_names(state.sts, state.namespace)}
+
+        return {}
+
+    def help(self, _: ReplState):
+        return f'{ExportTables.COMMAND} [* [in KEYSPACE]] | [TABLE] [as target-name] [with consistency <level>]\t export tables to Sqlite, Athena or CSV file'

adam/commands/export/export_databases.py

@@ -0,0 +1,245 @@
+from collections.abc import Callable
+from datetime import datetime
+import os
+import boto3
+
+from adam.commands.export.export_sessions import ExportSessions
+from adam.commands.export.importer import Importer
+from adam.config import Config
+from adam.repl_session import ReplSession
+from adam.repl_state import ReplState
+from adam.utils import debug, tabulize, log2, ing, log_exc
+from adam.utils_athena import Athena
+from adam.utils_sqlite import SQLite
+
+LIKE = 'e%_%'
+
+class ExportDatabases:
+    def run_query(query: str, database: str, show_query=False, output: Callable[[str], str] = None) -> int:
+        cnt: int = 0
+
+        if show_query:
+            log2(query)
+
+        log_file = None
+        if database.startswith('s'):
+            cnt0, log_file = SQLite.run_query(query, database=database, output=output)
+            cnt += cnt0
+        else:
+            cnt0, log_file = Athena.run_query(query, database=database, output=output)
+            cnt += cnt0
+
+        if Config().get('repl.history.push-cat-log-file', True):
+            if log_file and ReplSession().prompt_session:
+                ReplSession().prompt_session.history.append_string(f':sh cat {log_file}')
+
+        return cnt
+
+    def sessions_from_dbs(dbs: list[str]):
+
+        sessions = set()
+
+        for db in dbs:
+            sessions.add(db.split('_')[0])
+
+        return list(sessions)
+
+    def drop_export_dbs(db: str = None):
+        dbs: list[str] = []
+
+        if not db or db.startswith('s'):
+            dbs.extend(ExportDatabases.drop_sqlite_dbs(db))
+        if not db or db.startswith('e'):
+            dbs.extend(ExportDatabases.drop_athena_dbs(db))
+
+        return dbs
+
+    def drop_sqlite_dbs(db: str = None):
+        dbs = SQLite.database_names(db)
+        if dbs:
+            with ing(f'Droping {len(dbs)} SQLite databases'):
+                with log_exc():
+                    for db in dbs:
+                        file_path = f'{SQLite.local_db_dir()}/{db}'
+                        try:
+                            os.remove(file_path)
+                        except OSError as e:
+                            pass
+
+        return dbs
+
+    def drop_athena_dbs(db: str = None):
+        dbs = Athena.database_names(f'{db}_%' if db else LIKE)
+        if dbs:
+            with ing(f'Droping {len(dbs)} Athena databases'):
+                for db in dbs:
+                    query = f'DROP DATABASE {db} CASCADE'
+                    debug(query)
+                    Athena.query(query)
+
+        with ing(f'Deleting s3 folder: export'):
+            with log_exc():
+                if not db:
+                    db = ''
+
+                s3 = boto3.resource('s3')
+                bucket = s3.Bucket(Config().get('export.bucket', 'c3.ops--qing'))
+                bucket.objects.filter(Prefix=f'export/{db}').delete()
+
+        return dbs
+
+    def show_database(database: str):
+        if not database:
+            return
+
+        ExportDatabases.clear_cache()
+
+        keyspaces = {}
+        for table in ExportDatabases.table_names(database):
+            keyspace = table.split('.')[0]
+            if keyspace in keyspaces:
+                keyspaces[keyspace] += 1
+            else:
+                keyspaces[keyspace] = 1
+
+        tabulize(keyspaces.items(), lambda a: f'{a[0]},{a[1]}', header='SCHEMA,# of TABLES', separator=',')
+
+    def database_names():
+        return ExportDatabases.copy_database_names() + ExportDatabases.export_database_names()
+
+    def copy_database_names():
+        return list({n.split('_')[0] for n in SQLite.database_names()})
+
+    def export_database_names():
+        return list({n.split('_')[0] for n in Athena.database_names(LIKE)})
+
+    def database_names_with_keyspace_cnt(importer: str = None):
+        r = {}
+
+        names = []
+        if not importer:
+            names = SQLite.database_names() + Athena.database_names(LIKE)
+        elif importer == 'athena':
+            names = Athena.database_names(LIKE)
+        else:
+            names = SQLite.database_names()
+
+        for n in names:
+            tokens = n.split('_')
+            name = tokens[0]
+            keyspace = None
+            if len(tokens) > 1:
+                keyspace = tokens[1].replace('.db', '')
+
+            if keyspace == 'root':
+                continue
+
+            if name in r:
+                r[name] += 1
+            else:
+                r[name] = 1
+
+        return r
+
+    def table_names(session: str):
+        tables = []
+
+        for session in ExportDatabases._session_database_names(session):
+            if session.startswith('s'):
+                for table in SQLite.table_names(database=session):
+                    tables.append(f'{SQLite.keyspace(session)}.{table}')
+            else:
+                for table in Athena.table_names(database=session, function='export'):
+                    tables.append(f'{session}.{table}')
+
+        return tables
+
+    def _session_database_names(db: str):
+        eprefix = db
+        if '_' in db:
+            eprefix = db.split('_')[0]
+
+        if db.startswith('s'):
+            return SQLite.database_names(prefix=f'{eprefix}_')
+        else:
+            return Athena.database_names(like=f'{eprefix}_%')
+
+    def drop_databases(sts: str, pod: str, namespace: str, database: str = None):
+        importer = None
+        if database:
+            importer = Importer.importer_from_session(database)
+
+        sessions_done = ExportSessions.export_session_names(sts, pod, namespace, importer=importer, export_state='done')
+        sessions = ExportDatabases.sessions_from_dbs(ExportDatabases.drop_export_dbs(database))
+        if sessions_done and sessions:
+            intersects = list(set(sessions_done) & set(sessions))
+            with ing(f'Cleaning up {len(intersects)} completed sessions'):
+                ExportSessions.clean_up_sessions(sts, pod, namespace, list(intersects))
+                ExportSessions.clear_export_session_cache()
+
+    def clear_cache(database: str = None):
+        if not database or database.startswith('s'):
+            SQLite.clear_cache()
+        if not database or database.startswith('e'):
+            Athena.clear_cache()
+
+    def show_databases(importer: str = None):
+        lines = [f'{k}\t{v}' for k, v in ExportDatabases.database_names_with_keyspace_cnt(importer).items()]
+        tabulize(lines, header='NAME\tKEYSPACES', separator='\t')
+
+class ExportDatabaseService:
+    def __init__(self, handler: 'ExportDatabaseHandler'):
+        self.handler = handler
+
+    def sql(self, query: str, database: str = None, backgrounded = False):
+        if not database:
+            database = self.handler.state.export_session
+
+        def output(out: str):
+            log_prefix = Config().get('export.log-prefix', '/tmp/qing')
+            log_file = f'{log_prefix}-{datetime.now().strftime("%d%H%M%S")}-export.log'
+
+            with open(log_file, 'w') as f:
+                f.write(out)
+
+            return log_file
+
+        ExportDatabases.run_query(query, database, output = output if backgrounded else None, show_query = not backgrounded)
+
+    def drop(self, database: str):
+        state = self.handler.state
+
+        ExportDatabases.drop_databases(state.sts, state.pod, state.namespace, database)
+        ExportDatabases.clear_cache(database)
+        if state.export_session == database:
+            state.export_session = None
+
+    def drop_all(self):
+        state = self.handler.state
+
+        ExportDatabases.drop_databases(state.sts, state.pod, state.namespace)
+        ExportDatabases.clear_cache()
+
+        state.export_session = None
+
+    def show_databases(self, importer: str = None):
+        ExportDatabases.show_databases(importer)
+
+    def show_database(self, database: str = None):
+        if not database:
+            database = self.handler.state.export_session
+
+        ExportDatabases.show_database(database)
+
+class ExportDatabaseHandler:
+    def __init__(self, state: ReplState = None):
+        self.state = state
+
+    def __enter__(self):
+        return ExportDatabaseService(self)
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        return False
+
+def export_db(state: ReplState = None):
+    return ExportDatabaseHandler(state)

adam/commands/export/export_select.py

@@ -0,0 +1,59 @@
+from adam.commands import extract_trailing_options, validate_args
+from adam.commands.command import Command
+from adam.commands.export.export_databases import ExportDatabases, export_db
+from adam.repl_state import ReplState, RequiredState
+from adam.sql.sql_completer import SqlCompleter, SqlVariant
+from adam.utils_athena import Athena
+
+class ExportSelect(Command):
+    COMMAND = '.select'
+
+    # the singleton pattern
+    def __new__(cls, *args, **kwargs):
+        if not hasattr(cls, 'instance'): cls.instance = super(ExportSelect, cls).__new__(cls)
+
+        return cls.instance
+
+    def __init__(self, successor: Command=None):
+        super().__init__(successor)
+
+    def command(self):
+        return ExportSelect.COMMAND
+
+    def required(self):
+        return RequiredState.EXPORT_DB
+
+    def run(self, cmd: str, state: ReplState):
+        if not(args := self.args(cmd)):
+            return super().run(cmd, state)
+
+        with self.validate(args, state) as (args, state):
+            with extract_trailing_options(args, '&') as (args, backgrounded):
+                with validate_args(args, state, name='SQL statement') as query:
+                    with export_db(state) as dbs:
+                        dbs.sql(f'select {query}', backgrounded=backgrounded)
+
+                    return state
+
+    def completion(self, state: ReplState):
+        if not state.export_session:
+            return {}
+
+        db = state.export_session
+
+        # warm up the caches first time when x: drive is accessed
+        ExportDatabases.table_names(db)
+        Athena.column_names(database=db, function='export')
+        Athena.column_names(partition_cols_only=True, database=db, function='export')
+
+        return {ExportSelect.COMMAND: SqlCompleter(
+            lambda: ExportDatabases.table_names(db),
+            dml='select',
+            expandables={
+                'columns':lambda table: Athena.column_names(database=db, function='export'),
+            },
+            variant=SqlVariant.ATHENA
+        )}
+
+    def help(self, _: ReplState):
+        return f'.<sql-select-statements>\t run queries on export database'

adam/commands/export/export_select_x.py

@@ -0,0 +1,54 @@
+from adam.commands.command import Command
+from adam.commands.export.export_databases import ExportDatabases
+from adam.repl_state import ReplState, RequiredState
+from adam.sql.sql_completer import SqlCompleter, SqlVariant
+from adam.utils_athena import Athena
+
+# No action body, only for a help entry and auto-completion
+class ExportSelectX(Command):
+    COMMAND = 'select_on_x'
+
+    # the singleton pattern
+    def __new__(cls, *args, **kwargs):
+        if not hasattr(cls, 'instance'): cls.instance = super(ExportSelectX, cls).__new__(cls)
+
+        return cls.instance
+
+    def __init__(self, successor: Command=None):
+        super().__init__(successor)
+
+    def command(self):
+        return ExportSelectX.COMMAND
+
+    def required(self):
+        return RequiredState.EXPORT_DB
+
+    def completion(self, state: ReplState):
+        completions = {}
+
+        if state.device == ReplState.X:
+            completions = {'drop': SqlCompleter(
+                lambda: ExportDatabases.table_names(state.export_session),
+                dml='drop',
+                expandables={
+                    'export-dbs': lambda: ExportDatabases.database_names(),
+                    'columns':lambda _: Athena.column_names(database=state.export_session, function='export'),
+                },
+                variant=SqlVariant.ATHENA
+            )}
+
+            if state.export_session:
+                completions |= {dml: SqlCompleter(
+                    lambda: ExportDatabases.table_names(state.export_session),
+                    dml=dml,
+                    expandables={
+                        'export-dbs': lambda: ExportDatabases.database_names(),
+                        'columns':lambda _: Athena.column_names(database=state.export_session, function='export'),
+                    },
+                    variant=SqlVariant.ATHENA
+                ) for dml in ['select', 'preview']}
+
+        return completions
+
+    def help(self, _: ReplState):
+        return f'<sql-select-statements>\t run queries on export database'

adam/commands/export/export_sessions.py

@@ -0,0 +1,209 @@
+import functools
+import os
+import re
+
+from adam.commands.export.importer import Importer
+from adam.commands.export.utils_export import ExportTableStatus, csv_dir, find_files
+from adam.config import Config
+from adam.repl_state import ReplState
+from adam.utils import log2, tabulize, log, parallelize
+from adam.utils_k8s.cassandra_nodes import CassandraNodes
+from adam.utils_k8s.pods import Pods, log_prefix
+from adam.utils_k8s.statefulsets import StatefulSets
+from adam.utils_local import local_tmp_dir
+
+class ExportSessions:
+    def clear_export_session_cache():
+        ExportSessions.find_export_sessions.cache_clear()
+        ExportSessions.export_session_names.cache_clear()
+
+    @functools.lru_cache()
+    def export_session_names(sts: str, pod: str, namespace: str, importer: str = None, export_state = None):
+        if not sts or not namespace:
+            return []
+
+        if not pod:
+            pod = StatefulSets.pod_names(sts, namespace)[0]
+
+        if not pod:
+            return []
+
+        return [session for session, state in ExportSessions.find_export_sessions(pod, namespace, importer).items() if not export_state or state == export_state]
+
+    @functools.lru_cache()
+    def find_export_sessions(pod: str, namespace: str, importer: str = None, limit = 100):
+        sessions: dict[str, str] = {}
+
+        prefix = Importer.prefix_from_importer(importer)
+
+        log_files: list[str] = find_files(pod, namespace, f'{log_prefix()}-{prefix}*_*.log*')
+
+        if not log_files:
+            return {}
+
+        for log_file in log_files[:limit]:
+            m = re.match(f'{log_prefix()}-(.*?)_.*\.log?(.*)', log_file)
+            if m:
+                s = m.group(1)
+                state = m.group(2) # '', '.pending_import', '.done'
+                if state:
+                    state = state.strip('.')
+                else:
+                    state = 'in_export'
+
+                if s not in sessions:
+                    sessions[s] = state
+                elif sessions[s] == 'done' and state != 'done':
+                    sessions[s] = state
+
+        return sessions
+
+    def clean_up_all_sessions(sts: str, pod: str, namespace: str):
+        if not sts or not namespace:
+            return False
+
+        if not pod:
+            pod = StatefulSets.pod_names(sts, namespace)[0]
+
+        CassandraNodes.exec(pod, namespace, f'rm -rf {csv_dir()}/*', show_out=Config().is_debug(), shell='bash')
+        CassandraNodes.exec(pod, namespace, f'rm -rf {log_prefix()}-*.log*', show_out=Config().is_debug(), shell='bash')
+
+        return True
+
+    def clean_up_sessions(sts: str, pod: str, namespace: str, sessions: list[str], max_workers = 0):
+        if not sessions:
+            return []
+
+        if not max_workers:
+            max_workers = Config().action_workers('export', 8)
+
+        with parallelize(sessions,
+                         max_workers,
+                         msg='Cleaning|Cleaned up {size} export sessions') as exec:
+            cnt_tuples = exec.map(lambda session: ExportSessions.clean_up_session(sts, pod, namespace, session, True))
+            csv_cnt = 0
+            log_cnt = 0
+            for (csv, log) in cnt_tuples:
+                csv_cnt += csv
+                log_cnt += log
+
+            return csv_cnt, log_cnt
+
+    def clean_up_session(sts: str, pod: str, namespace: str, session: str, multi_tables = True):
+        if not sts or not namespace:
+            return 0, 0
+
+        if not pod:
+            pod = StatefulSets.pod_names(sts, namespace)[0]
+
+        if not pod:
+            return 0, 0
+
+        csv_cnt = 0
+        log_cnt = 0
+
+        log_files: list[str] = find_files(pod, namespace, f'{log_prefix()}-{session}_*.log*')
+
+        for log_file in log_files:
+            m = re.match(f'{log_prefix()}-{session}_(.*?)\.(.*?)\.log.*', log_file)
+            if m:
+                table = m.group(2)
+
+                CassandraNodes.exec(pod, namespace, f'rm -rf {csv_dir()}/{session}_{table}', show_out=not multi_tables, shell='bash')
+                csv_cnt += 1
+
+                CassandraNodes.exec(pod, namespace, f'rm -rf {log_file}', show_out=not multi_tables, shell='bash')
+                log_cnt += 1
+
+        return csv_cnt, log_cnt
+
+    def show_session(sts: str, pod: str, namespace: str, session: str):
+        if not pod:
+            pod = StatefulSets.pod_names(sts, namespace)[0]
+
+        if not pod:
+            return
+
+        tables, _ = ExportTableStatus.from_session(sts, pod, namespace, session)
+        log()
+        tabulize(tables,
+                 lambda t: f'{t.keyspace}\t{t.target_table}\t{"export_completed_pending_import" if t.status == "pending_import" else t.status}\t{t.csv_file}',
+                 header='KEYSPACE\tTARGET_TABLE\tSTATUS\tCSV_FILES',
+                 separator='\t')
+
+    def download_session(sts: str, pod: str, namespace: str, session: str):
+        if not pod:
+            pod = StatefulSets.pod_names(sts, namespace)[0]
+
+        if not pod:
+            return
+
+        tables, _ = ExportTableStatus.from_session(sts, pod, namespace, session)
+        def download_csv(table):
+            from_path: str = table.csv_file
+
+            to_path = from_path.replace(csv_dir(), local_tmp_dir())
+            os.makedirs(os.path.dirname(to_path), exist_ok=True)
+            Pods.download_file(pod, 'cassandra', namespace, from_path, to_path)
+
+            log2(f'[{session}] Downloaded to {to_path}.')
+
+        with parallelize(tables,
+                         workers=Config().get('download.workers', 8),
+                         msg='Downloading|Downloaded {size} csv files') as exec:
+            exec.map(download_csv)
+
+class ExportSessionService:
+    def __init__(self, handler: 'ExportSessionHandler'):
+        self.handler = handler
+
+    def clean_up(self, sessions: list[str]):
+        state = self.handler.state
+
+        csv_cnt, log_cnt = ExportSessions.clean_up_sessions(state.sts, self.pod(), state.namespace, sessions)
+
+        log(f'Removed {csv_cnt} csv and {log_cnt} log files.')
+
+        ExportSessions.clear_export_session_cache()
+
+    def clean_up_all(self):
+        state = self.handler.state
+
+        if ExportSessions.clean_up_all_sessions(state.sts, self.pod(), state.namespace):
+            ExportSessions.clear_export_session_cache()
+
+    def show_all_sessions(self):
+        state = self.handler.state
+
+        sessions = sorted(ExportSessions.find_export_sessions(self.pod(), state.namespace).items(), reverse=True)
+        tabulize(sessions, lambda args: f'{args[0]}\t{args[1]}', header='EXPORT_SESSION\tSTATUS', separator='\t')
+
+    def show_session(self, session: str):
+        state = self.handler.state
+        ExportSessions.show_session(state.sts, self.pod(), state.namespace, session)
+
+    def download_session(self, session: str):
+        state = self.handler.state
+        ExportSessions.download_session(state.sts, self.pod(), state.namespace, session)
+
+    def pod(self):
+        state = self.handler.state
+
+        pod = state.pod
+        if not pod:
+            pod = StatefulSets.pod_names(state.sts, state.namespace)[0]
+
+        return pod
+
+class ExportSessionHandler:
+    def __init__(self, state: ReplState = None):
+        self.state = state
+
+    def __enter__(self):
+        return ExportSessionService(self)
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        return False
+
+def export_session(state: ReplState = None):
+    return ExportSessionHandler(state)