Removed code duplication while importing datafiles

mathesar/imports/base.py

@@ -0,0 +1,36 @@
+from mathesar.database.base import create_mathesar_engine
+from mathesar.models.base import Table
+from mathesar.imports.csv import create_db_table_from_csv_data_file
+from mathesar.imports.json import create_db_table_from_json_data_file
+from db.tables.operations.select import get_oid_from_table
+from mathesar.errors import InvalidTableError
+
+ALLOWED_DELIMITERS = ",\t:|;"
+SAMPLE_SIZE = 20000
+CHECK_ROWS = 10
+
+
+def create_table_from_data_file(data_file, name, schema, comment=None):
+    engine = create_mathesar_engine(schema.database.name)
+    if data_file.type == 'csv' or data_file.type == 'tsv':
+        db_table = create_db_table_from_csv_data_file(
+            data_file, name, schema, comment=comment
+        )
+    elif data_file.type == 'json':
+        db_table = create_db_table_from_json_data_file(
+            data_file, name, schema, comment=comment
+        )
+    else:
+        raise InvalidTableError
+    db_table_oid = get_oid_from_table(db_table.name, db_table.schema, engine)
+    # Using current_objects to create the table instead of objects. objects
+    # triggers re-reflection, which will cause a race condition to create the table
+    table = Table.current_objects.get(
+        oid=db_table_oid,
+        schema=schema,
+    )
+    table.import_verified = False
+    table.save()
+    data_file.table_imported_to = table
+    data_file.save()
+    return table

mathesar/imports/csv.py

@@ -2,15 +2,13 @@
 
 import clevercsv as csv
 
-from db.identifiers import truncate_if_necessary
 from db.tables.operations.alter import update_pk_sequence_to_latest
 from mathesar.database.base import create_mathesar_engine
-from mathesar.models.base import Table
 from db.records.operations.insert import insert_records_from_csv
 from db.tables.operations.create import create_string_column_table
-from db.tables.operations.select import get_oid_from_table
 from db.tables.operations.drop import drop_table
 from mathesar.errors import InvalidTableError
+from mathesar.imports.utils import process_column_names
 from db.constants import ID, ID_ORIGINAL, COLUMN_NAME_TEMPLATE
 from psycopg2.errors import IntegrityError, DataError
 
@@ -109,98 +107,50 @@ def get_sv_reader(file, header, dialect=None):
     return reader
 
 
-def create_db_table_from_data_file(data_file, name, schema, comment=None):
+def insert_data_from_csv_data_file(name, schema, column_names, engine, comment, data_file):
+    dialect = csv.dialect.SimpleDialect(data_file.delimiter, data_file.quotechar,
+                                        data_file.escapechar)
+    encoding = get_file_encoding(data_file.file)
+    table = create_string_column_table(
+        name=name,
+        schema=schema.name,
+        column_names=column_names,
+        engine=engine,
+        comment=comment,
+    )
+    insert_records_from_csv(
+        table,
+        engine,
+        data_file.file.path,
+        column_names,
+        data_file.header,
+        delimiter=dialect.delimiter,
+        escape=dialect.escapechar,
+        quote=dialect.quotechar,
+        encoding=encoding
+    )
+    return table
+
+
+def create_db_table_from_csv_data_file(data_file, name, schema, comment=None):
     db_name = schema.database.name
     engine = create_mathesar_engine(db_name)
     sv_filename = data_file.file.path
     header = data_file.header
     dialect = csv.dialect.SimpleDialect(data_file.delimiter, data_file.quotechar,
                                         data_file.escapechar)
-    encoding = get_file_encoding(data_file.file)
     with open(sv_filename, 'rb') as sv_file:
         sv_reader = get_sv_reader(sv_file, header, dialect=dialect)
-        column_names = _process_column_names(sv_reader.fieldnames)
-        table = create_string_column_table(
-            name=name,
-            schema=schema.name,
-            column_names=column_names,
-            engine=engine,
-            comment=comment,
-        )
+        column_names = process_column_names(sv_reader.fieldnames)
     try:
-        insert_records_from_csv(
-            table,
-            engine,
-            sv_filename,
-            column_names,
-            header,
-            delimiter=dialect.delimiter,
-            escape=dialect.escapechar,
-            quote=dialect.quotechar,
-            encoding=encoding
-        )
+        table = insert_data_from_csv_data_file(name, schema, column_names, engine, comment, data_file)
         update_pk_sequence_to_latest(engine, table)
     except (IntegrityError, DataError):
         drop_table(name=name, schema=schema.name, engine=engine)
         column_names_alt = [
             column_name if column_name != ID else ID_ORIGINAL
             for column_name in column_names
         ]
-        table = create_string_column_table(
-            name=name,
-            schema=schema.name,
-            column_names=column_names_alt,
-            engine=engine,
-            comment=comment,
-        )
-        insert_records_from_csv(
-            table,
-            engine,
-            sv_filename,
-            column_names_alt,
-            header,
-            delimiter=dialect.delimiter,
-            escape=dialect.escapechar,
-            quote=dialect.quotechar,
-            encoding=encoding
-        )
+        insert_data_from_csv_data_file(name, schema, column_names_alt, engine, comment, data_file)
     reset_reflection(db_name=db_name)
     return table
-
-
-def _process_column_names(column_names):
-    column_names = (
-        column_name.strip()
-        for column_name
-        in column_names
-    )
-    column_names = (
-        truncate_if_necessary(column_name)
-        for column_name
-        in column_names
-    )
-    column_names = (
-        f"{COLUMN_NAME_TEMPLATE}{i}" if name == '' else name
-        for i, name
-        in enumerate(column_names)
-    )
-    return list(column_names)
-
-
-def create_table_from_csv(data_file, name, schema, comment=None):
-    engine = create_mathesar_engine(schema.database.name)
-    db_table = create_db_table_from_data_file(
-        data_file, name, schema, comment=comment
-    )
-    db_table_oid = get_oid_from_table(db_table.name, db_table.schema, engine)
-    # Using current_objects to create the table instead of objects. objects
-    # triggers re-reflection, which will cause a race condition to create the table
-    table = Table.current_objects.get(
-        oid=db_table_oid,
-        schema=schema,
-    )
-    table.import_verified = False
-    table.save()
-    data_file.table_imported_to = table
-    data_file.save()
-    return table

mathesar/imports/json.py

@@ -1,14 +1,12 @@
 import json
 
-from db.identifiers import truncate_if_necessary
 from db.tables.operations.alter import update_pk_sequence_to_latest
 from mathesar.database.base import create_mathesar_engine
-from mathesar.models.base import Table
 from db.records.operations.insert import insert_records_from_json
 from db.tables.operations.create import create_string_column_table
-from db.tables.operations.select import get_oid_from_table
 from db.tables.operations.drop import drop_table
-from db.constants import ID, ID_ORIGINAL, COLUMN_NAME_TEMPLATE
+from mathesar.imports.utils import process_column_names
+from db.constants import ID, ID_ORIGINAL
 from psycopg2.errors import IntegrityError, DataError
 
 from mathesar.state import reset_reflection
@@ -23,82 +21,39 @@ def get_column_names_from_json(data_file):
     return list(data.keys())
 
 
-def _process_column_names(column_names):
-    column_names = (
-        column_name.strip()
-        for column_name
-        in column_names
-    )
-    column_names = (
-        truncate_if_necessary(column_name)
-        for column_name
-        in column_names
+def insert_data_from_json_data_file(name, schema, column_names, engine, comment, json_filepath):
+    table = create_string_column_table(
+        name=name,
+        schema=schema.name,
+        column_names=column_names,
+        engine=engine,
+        comment=comment,
     )
-    column_names = (
-        f"{COLUMN_NAME_TEMPLATE}{i}" if name == '' else name
-        for i, name
-        in enumerate(column_names)
+    insert_records_from_json(
+        table,
+        engine,
+        json_filepath
     )
-    return list(column_names)
+    return table
 
 
 def create_db_table_from_json_data_file(data_file, name, schema, comment=None):
     db_name = schema.database.name
     engine = create_mathesar_engine(db_name)
     json_filepath = data_file.file.path
-    column_names = _process_column_names(
-        get_column_names_from_json(data_file.file.path)
-    )
-    table = create_string_column_table(
-        name=name,
-        schema=schema.name,
-        column_names=column_names,
-        engine=engine,
-        comment=comment,
+    column_names = process_column_names(
+        get_column_names_from_json(json_filepath)
     )
     try:
-        insert_records_from_json(
-            table,
-            engine,
-            json_filepath
-        )
+        table = insert_data_from_json_data_file(name, schema, column_names, engine, comment, json_filepath)
         update_pk_sequence_to_latest(engine, table)
     except (IntegrityError, DataError):
         drop_table(name=name, schema=schema.name, engine=engine)
         column_names_alt = [
             fieldname if fieldname != ID else ID_ORIGINAL
             for fieldname in column_names
         ]
-        table = create_string_column_table(
-            name=name,
-            schema=schema.name,
-            column_names=column_names_alt,
-            engine=engine,
-            comment=comment,
-        )
-        insert_records_from_json(
-            table,
-            engine,
-            json_filepath
-        )
-    reset_reflection(db_name=db_name)
-    return table
+        table = insert_data_from_json_data_file(name, schema, column_names_alt, engine, comment, json_filepath)
 
-
-def create_table_from_json(data_file, name, schema, comment=None):
-    engine = create_mathesar_engine(schema.database.name)
-    db_table = create_db_table_from_json_data_file(
-        data_file, name, schema, comment=comment
-    )
-    db_table_oid = get_oid_from_table(
-        db_table.name, db_table.schema, engine
-    )
-    table = Table.current_objects.get(
-        oid=db_table_oid,
-        schema=schema,
-    )
-    table.import_verified = False
-    table.save()
-    data_file.table_imported_to = table
-    data_file.save()
+    reset_reflection(db_name=db_name)
     return table

mathesar/imports/utils.py

@@ -0,0 +1,21 @@
+from db.identifiers import truncate_if_necessary
+from db.constants import COLUMN_NAME_TEMPLATE
+
+
+def process_column_names(column_names):
+    column_names = (
+        column_name.strip()
+        for column_name
+        in column_names
+    )
+    column_names = (
+        truncate_if_necessary(column_name)
+        for column_name
+        in column_names
+    )
+    column_names = (
+        f"{COLUMN_NAME_TEMPLATE}{i}" if name == '' else name
+        for i, name
+        in enumerate(column_names)
+    )
+    return list(column_names)

mathesar/tests/conftest.py

@@ -20,7 +20,7 @@
 from db.schemas.utils import get_schema_oid_from_name
 
 import mathesar.tests.conftest
-from mathesar.imports.csv import create_table_from_csv
+from mathesar.imports.base import create_table_from_data_file
 from mathesar.models.base import Schema, Table, Database, DataFile
 from mathesar.models.base import Column as mathesar_model_column
 from mathesar.models.users import DatabaseRole, SchemaRole, User
@@ -322,7 +322,7 @@ def create_table(create_schema):
     def _create_table(table_name, schema_name, csv_filepath):
         data_file = _get_datafile_for_path(csv_filepath)
         schema_model = create_schema(schema_name)
-        return create_table_from_csv(data_file, table_name, schema_model)
+        return create_table_from_data_file(data_file, table_name, schema_model)
     return _create_table