dataforge-core 0.1.1__py3-none-any.whl

dataforge/importProject.py

@@ -0,0 +1,128 @@
+import json
+import os
+import sys
+
+import yaml
+from .mainConfig import MainConfig
+from .miniSparky import MiniSparky
+
+
+class ImportProject:
+    def __init__(self, config: MainConfig):
+        self.ms = None
+        self._config = config
+        self.import_id = 0
+
+    def start(self):
+        _import_id = self._config.pg.sql("select meta.svc_import_start()")
+        self.import_id = int(_import_id)
+        print('Started import with id ', self.import_id)
+
+    def validate(self):
+        print(f"Validating project path {self._config.source_path}")
+        meta_flag = False
+        source_flag = False
+        with os.scandir(self._config.source_path) as entries:
+            for file in entries:
+                if file.is_dir() and file.name == 'sources':
+                    source_flag = True
+                elif file.name == "meta.yaml":
+                    meta_flag = True
+        if not meta_flag:
+            print(f"Missing meta.yaml in project path {self._config.source_path}")
+            sys.exit(1)
+        if not source_flag:
+            print(f"Missing sources folder in project path {self._config.source_path}")
+            sys.exit(1)
+
+    def load(self):
+        self.validate()
+        self.start()
+        with os.scandir(self._config.source_path) as entries:
+            for file in entries:
+                if file.is_dir() and file.name in ('sources', 'outputs'):
+                    self.list_files(file.path, file.name)
+                elif file.name.endswith(".yaml"):
+                    self.load_file(file.path, file.name)
+        self._config.pg.sql("SELECT meta.svc_import_complete(%s, 'I')", [self.import_id])
+        self._config.pg.sql("SELECT meta.imp_parse_objects(%s)", [self.import_id])
+        print("Files parsed")
+        self._config.pg.sql("SELECT meta.svc_import_execute(%s)", [self.import_id])
+        print("Objects loaded")
+        self.test_expressions()
+        print("Expressions validated")
+        print("Import completed successfully")
+        self.write_log()
+        self.write_queries("source")
+        self.write_queries("output")
+
+    def list_files(self, path: str, folder_name: str):
+        print("Importing files..")
+        with os.scandir(path) as entries:
+            for file in entries:
+                if file.is_file() & file.name.endswith(".yaml"):
+                    self.load_file(file.path, folder_name + '/' + file.name)
+
+    def load_file(self, full_path: str, path: str):
+        print(path)
+        with open(full_path, 'r') as file:
+            file_js = yaml.safe_load(file)
+            self._config.pg.sql("SELECT meta.svc_import_load_object(%s, %s, %s)",
+                                (self.import_id, path, json.dumps(file_js)))
+
+    def test_expressions(self):
+        exps = self._config.pg.sql("SELECT meta.impc_test_expressions(%s)", [self.import_id])
+        self.ms = MiniSparky(self._config)
+        self.test_expressions_recursive(exps)
+        self.ms.stop()
+        del self.ms
+
+    def test_expressions_recursive(self, test_expressions, recursion_level=0):
+        try:
+            test_results = []
+            for exp in test_expressions:
+                query_result = self.ms.execute_query(exp['expression'])
+                del exp['expression']
+                exp['result'] = query_result
+                test_results.append(exp)
+            #  update test results
+            test_results_str = json.dumps(test_results)
+            res = self._config.pg.sql("SELECT meta.impc_update_test_results(%s, %s)",
+                                      (self.import_id, test_results_str))
+            if res.get('error'):
+                self.fail_import('Invalid expression detected. See log file for details')
+            if recursion_level > 20:
+                self.fail_import('Maximum recursion exceeded while testing expressions. Check error logs and '
+                                 'expression test tables for more details')
+            if len(res['next']) > 0:
+                self.test_expressions_recursive(res['next'], recursion_level + 1)
+        except Exception as e:
+            print(e)
+            self.fail_import(str(e))
+
+    def fail_import(self, message):
+        print(f"Import failed: {message}")
+        self._config.pg.sql("SELECT meta.svc_import_complete(%s, 'F', %s)", (self.import_id, message))
+        self.write_log()
+        sys.exit(1)
+
+    def write_log(self):
+        log_file = self._config.pg.sql("SELECT meta.svc_import_get_log(%s)", [self.import_id])
+        with open(self._config.log_path, "w") as file:
+            # Write the string to the file
+            file.write(log_file)
+
+    def write_queries(self, out_type: str):
+        queries = self._config.pg.sql(f"select meta.svcc_get_{out_type}_queries(%s)", [self.import_id])
+        if not queries:
+            return
+        if out_type == "source":
+            out_path = self._config.output_source_path
+        elif out_type == "output":
+            out_path = self._config.output_output_path
+        for query in queries:
+            file_name = os.path.join(out_path, query['file_name'])
+            with open(file_name, "w") as file:
+                # Write the string to the file
+                file.write(query['query'])
+        print(f"Generated {len(queries)} {out_type} queries")

dataforge/main.py

@@ -0,0 +1,12 @@
+from .importProject import ImportProject
+from .mainConfig import MainConfig
+
+
+def main():
+    conf = MainConfig()
+    imp = ImportProject(conf)
+    imp.load()
+
+
+if __name__ == '__main__':
+    main()

dataforge/mainConfig.py

@@ -0,0 +1,72 @@
+import argparse
+import os
+import shutil
+import sys
+import traceback
+from types import ModuleType
+
+from importlib_resources.abc import Traversable
+from .pg import Pg
+import importlib_resources
+
+
+class MainConfig:
+    def __init__(self):
+        # check if Java is installed
+        if os.environ.get('JAVA_HOME') is None:
+            print("Java is not installed or JAVA_HOME environment variable is not set")
+            sys.exit(1)
+        # parse command line args
+        _parser = argparse.ArgumentParser(
+            prog='dataforge core',
+            description='Dataforge Core compiles project and generates source and output SQL queries',
+            epilog='Try our cloud product')
+        _parser.add_argument('source', type=str, help='Project folder', metavar='<Project Path>', nargs='?')
+        _parser.add_argument('--init', '-i', action='store_true', help='Initialize project folder')
+        _parser.add_argument('--seed', action='store_true', help='Deploy and seed postgres database')
+        _parser.add_argument('--connect', '-c', type=str, help='Connect to postgres database',
+                             metavar='<Postgres connection string>')
+
+        args = _parser.parse_args()
+        if args.connect:
+            self.pg = Pg(args.connect, initialize=True)
+            self.pg.seed()
+            sys.exit(0)
+        else:
+            self.pg = Pg()
+        if args.seed:
+            self.pg.seed()
+        self.source_path = os.getcwd() if args.source is None else args.source
+        if args.init:
+            try:
+                if self.pg.confirm_action(f"All files and subfolders in {self.source_path} will be deleted. Continue (y/n)? "):
+                    shutil.rmtree(self.source_path, ignore_errors=True)
+                    os.makedirs(self.source_path)
+                    self.traverse_resource_dir(importlib_resources.files().joinpath('resources', 'project'))
+                print(f"Initialized project in {self.source_path}")
+            except Exception as e:
+                traceback.print_stack()
+                print(f"Error initializing project in {self.source_path} : {e}")
+            sys.exit(0)
+        self.output_path = os.path.join(self.source_path, 'target')
+        self.log_path = os.path.join(self.output_path, 'log.txt')
+        shutil.rmtree(self.output_path, ignore_errors=True)
+        os.makedirs(self.output_path)
+        self.output_source_path = os.path.join(self.output_path, 'sources')
+        os.makedirs(self.output_source_path)
+        self.output_output_path = os.path.join(self.output_path, 'outputs')
+        os.makedirs(self.output_output_path)
+
+    def traverse_resource_dir(self, resource: Traversable, folder=''):
+        for file in resource.iterdir():
+            if file.is_dir():
+                os.makedirs(os.path.join(self.source_path, folder, file.name))
+                self.traverse_resource_dir(resource.joinpath(file.name), file.name)
+            if file.is_file():
+                self.copy_resource_file(folder, file)
+
+    def copy_resource_file(self, folder: str, resource: Traversable):
+        file_name = os.path.join(self.source_path, folder, resource.name)
+        with open(file_name, "w") as file:
+            # Write the string to the file
+            file.write(resource.read_text())

dataforge/miniSparky.py

@@ -0,0 +1,83 @@
+from pyspark.sql import SparkSession
+from pyspark.sql.types import DataType, StructType, ArrayType, DecimalType
+from .mainConfig import MainConfig
+
+
+class MiniSparky:
+    def __init__(self, config: MainConfig):
+        self._config = config
+        self.spark = SparkSession.builder \
+            .appName("dfCore") \
+            .master("local[1]") \
+            .config("spark.driver.extraJavaOptions", "-Dlog4j.logger.level=FATAL") \
+            .config("spark.executor.extraJavaOptions", "-Dlog4j.logger.level=FATAL") \
+            .config("spark.log.level", "FATAL") \
+            .config("spark.driver.memory", "512m") \
+            .config("spark.executor.memory", "512m") \
+            .config("spark.executor.cores", "1") \
+            .config("spark.ui.enabled", "false") \
+            .getOrCreate()
+
+        self.spark.sql(
+            """SELECT CAST(-87.68 as DECIMAL(10,2)) `decimal` , CAST(13518864 as BIGINT) `bigint`, CAST('Western Ave & Walton St' as STRING) `string`, CAST(130 AS INT) `int`, CAST(130 AS INT) `integer`, CAST(41.90331 as FLOAT) `float` , CAST(87.67695 as DOUBLE) `double`, CAST('2017-03-31' as DATE) `date`, CAST('2017-03-31T23:19:17.000+0000' as TIMESTAMP) `timestamp` , true `boolean`, CAST(9999999999 as BIGINT) `long`
+                UNION ALL
+                SELECT CAST(-8127.68 as DECIMAL(10,2)) `decimal` , CAST(1518864 as BIGINT) `bigint`, CAST('Western Ave & Walton St' as STRING) `string`, CAST(130 AS INT) `int`, CAST(130 AS INT) `integer`, CAST(41.90331 as FLOAT) `float` , CAST(87.67695 as DOUBLE) `double`, CAST('2020-03-31' as DATE) `date`, CAST('2020-03-31T23:19:17' as TIMESTAMP) `timestamp` , true `boolean`, CAST(99999999991 as BIGINT) `long`
+                """).createOrReplaceTempView("datatypes")
+
+        _res = self._config.pg.sql("select meta.svc_select_attribute_types_spark_to_hive()")
+        self.type_map = {}
+        for x in _res:
+            self.type_map[x['spark_type']] = x['hive_type']
+
+    @staticmethod
+    def get_spark_type(spark_type: DataType) -> str:
+        match spark_type:
+            case StructType(_):
+                return "StructType"
+            case ArrayType(_, _):
+                return "ArrayType"
+            case DecimalType():
+                return "DecimalType"
+            case _:
+                return str(spark_type).rstrip("()")
+
+    def execute_query(self, query: str):
+        try:
+            df = self.spark.sql(query)
+            field = df.schema.fields[0]
+            is_null = df.head()[0] is None
+            col_name = field.name
+            dt = field.dataType
+            spark_type = self.get_spark_type(dt)
+            att_schema = field.dataType.json()
+            data_type = self.type_map.get(spark_type)
+
+            if is_null:
+                return {
+                    "type": "warning",
+                    "data_type": data_type,
+                    "att_schema": att_schema,
+                    "message": "NULL values detected! This typically indicates improper type casting or that you're doing some very complex logic"
+                }
+            elif col_name != "col1":
+                return {
+                    "type": "error",
+                    "message": "Extraneous input detected at end of expression"
+                }
+            else:
+                return {
+                    "type": "success",
+                    "data_type": data_type,
+                    "att_schema": att_schema
+                }
+        except Exception as e:
+            print("spark exception, printing stack trace")
+            return {
+                "type": "error",
+                "message": str(e)
+            }
+
+    def stop(self):
+        self.spark.sparkContext.stop()
+        del self.spark
+

dataforge/pg.py

@@ -0,0 +1,70 @@
+import os
+import sys
+import psycopg2
+from importlib_resources import files
+
+
+class Pg:
+    def __init__(self, connection_string: str = None, initialize=False):
+        try:
+            if initialize:
+                self.initialize(connection_string)
+            else:
+                conn_string = os.environ.get('DATAFORGE_PG_CONNECTION')
+                print(f"Connecting to Postgres..")
+                if conn_string is None:
+                    print("Postgres connection is not initialized. Run with --mode init")
+                    sys.exit(1)
+                self.conn = psycopg2.connect(conn_string)
+                self.conn.set_session(autocommit=True)
+
+        except Exception as e:
+            print(f"Error connecting to Postgres: {e}")
+            sys.exit(1)
+
+    def sql(self, query: str, params=None, fetch=True):
+        # Execute a query
+        cur = self.conn.cursor()
+        cur.execute(query, params)
+        # Retrieve query results
+        res = cur.fetchone() if fetch else [None]
+        cur.close()
+        return res[0]
+
+    def initialize(self, connection_string: str):
+        # Execute a query
+        try:
+            print("Platform :", sys.platform)
+            self.conn = psycopg2.connect(connection_string)
+            self.sql("select 1")  # execute test query
+            match sys.platform:
+                case 'win32' | 'cygwin':
+                    os.system(f"SETX DATAFORGE_PG_CONNECTION \"{connection_string}\"")
+                case _:
+                    os.system(f"export DATAFORGE_PG_CONNECTION=\"{connection_string}\"")
+            # Change connection
+            print("Please restart your console")
+        except Exception as e:
+            print(f"Error initializing Postgres database or insufficient permissions. Details: {e}")
+            sys.exit(1)
+
+    def seed(self):
+        schemas = self.sql(
+            "select string_agg(schema_name,',') from information_schema.schemata where schema_name IN ('meta','log')")
+        if schemas:
+            if not self.confirm_action(
+                    f"All objects in schema(s) {schemas} in postgres database will be deleted. Do you want to continue (y/n)?"):
+                sys.exit(1)
+        #  Drop schemas
+        self.sql("DROP SCHEMA IF EXISTS meta CASCADE;"
+                 "DROP SCHEMA IF EXISTS log CASCADE;", fetch=False)
+        #  Deploy DB code
+        print("Initializing database..")
+        deploy_sql = files().joinpath('resources', 'pg_deploy.sql').read_text()
+        self.sql(deploy_sql, fetch=False)
+        print("Database initialized")
+
+    @staticmethod
+    def confirm_action(message: str):
+        confirmation = input(message).strip().lower()
+        return confirmation in ('yes', 'y')