Merge pull request #97 from MartinXPN/feature/sql

Feature/sql

.github/workflows/pre-commit.yaml

@@ -11,4 +11,6 @@ jobs:
     steps:
     - uses: actions/checkout@v4
     - uses: actions/setup-python@v4
+      with:
+        python-version: 3.12
     - uses: pre-commit/[email protected]

bouncer/coderunners.py

@@ -29,6 +29,8 @@ def from_language(language: str) -> 'CodeRunner':
             return JsRunner()
         if language in JavaRunner.supported_standards:
             return JavaRunner()
+        if language in SQLiteRunner.supported_standards:
+            return SQLiteRunner()
         raise ValueError(f'{language} does not have a compiler yet')
 
     def invoke(self, aws_lambda_client, request: SubmissionRequest) -> SubmissionResult:
@@ -100,3 +102,12 @@ class JavaRunner(CodeRunner):
     @property
     def name(self) -> str:
         return 'CodeRunnerJava'
+
+
+@dataclass
+class SQLiteRunner(CodeRunner):
+    supported_standards = {'sql', 'sqlite'}
+
+    @property
+    def name(self) -> str:
+        return 'CodeRunnerSQLite'

coderunners/compilers.py

@@ -5,7 +5,7 @@
 from pathlib import Path
 from typing import ClassVar
 
-from coderunners.executors import Executor, ProcessExecutor
+from coderunners.executors import Executor, ProcessExecutor, SQLiteExecutor
 from coderunners.process import Process
 from models import RunResult, Status
 
@@ -24,7 +24,7 @@ def find_main_file_path(cls, submission_paths: list[Path], main_file_name: str)
 
     @staticmethod
     def from_language(language: str) -> 'Compiler':
-        language = language.lower()
+        language = language.lower().strip()
         if language in TxtCompiler.supported_standards:
             return TxtCompiler()
         if language in CppCompiler.supported_standards:
@@ -39,6 +39,8 @@ def from_language(language: str) -> 'Compiler':
             return JsCompiler(language_standard=language)
         if language in JavaCompiler.supported_standards:
             return JavaCompiler()
+        if language in SQLiteCompiler.supported_standards:
+            return SQLiteCompiler()
         raise ValueError(f'{language} does not have a compiler yet')
 
 
@@ -186,3 +188,20 @@ def compile(self, submission_paths: list[Path]):
         compile_res = Process(f'cd {self.build_dir} && jar cvf Main.jar *', timeout=10, memory_limit_mb=512).run()
         print('Compile res:', compile_res)
         return ProcessExecutor(command=command), compile_res
+
+
+@dataclass
+class SQLiteCompiler(Compiler):
+    supported_standards = {'sql', 'sqlite'}
+    db_name: str = 'main.db'
+
+    def compile(self, submission_paths: list[Path]):
+        if len(submission_paths) != 1:
+            return ProcessExecutor(command='echo "Only one file is allowed"'), RunResult(
+                status=Status.CE, memory=0, time=0, return_code=0, outputs=None,
+                errors='Only one file is allowed for SQL submissions',
+            )
+
+        script = submission_paths[0].read_text()
+        executor = SQLiteExecutor(script=script, db_name=self.db_name)
+        return executor, RunResult(status=Status.OK, memory=0, time=0, return_code=0, outputs=None, errors=None)

coderunners/executors.py

@@ -1,10 +1,12 @@
+import sqlite3
 from abc import ABC, abstractmethod
 from collections.abc import Iterable
 from dataclasses import dataclass
+from io import StringIO
 from pathlib import Path
 
 from coderunners.process import Process
-from models import RunResult, TestCase
+from models import RunResult, Status, TestCase
 
 
 class Executor(ABC):
@@ -58,3 +60,104 @@ def cleanup(self, test: TestCase) -> None:
             if (self.ROOT / filename).exists():
                 print('Removing file at:', self.ROOT / filename)
                 (self.ROOT / filename).unlink()
+
+
+@dataclass
+class SQLiteExecutor(Executor):
+    script: str
+    db_name: str = 'main.db'
+
+    def __post_init__(self):
+        self.db = sqlite3.connect(self.db_name)
+
+    def __del__(self):
+        self.db.close()
+
+    def run(self, test: TestCase, **kwargs) -> RunResult:
+        """
+        self.script is the SQL script that needs to be run on the database
+        test.input is the initialization SQL script
+        test.input_files are all the tables that need to be populated
+        test.target is the expected output of the SQL script (can be empty)
+        test.target_files are all the tables that need to be populated by the SQL script
+        """
+        import pandas as pd
+        cursor = self.db.cursor()
+
+        try:
+            cursor.executescript(test.input)
+            self.db.commit()
+        except sqlite3.Error as e:
+            cursor.close()
+            return RunResult(
+                status=Status.RUNTIME_ERROR, memory=0, time=0, return_code=0, outputs=None,
+                errors=str(e),
+            )
+
+        for filename, content in (test.input_files or {}).items():
+            # Load the content of the file into a dataframe and then load it into the db (filename)
+            try:
+                print('Creating table:', filename)
+                csv_data = StringIO(content)
+                df = pd.read_csv(csv_data)
+                print(df.head())
+                df.to_sql(filename, self.db, if_exists='replace', index=False)
+                print('--- Done ---')
+            except (sqlite3.Error, pd.errors.ParserError, pd.errors.DatabaseError, ValueError) as e:
+                cursor.close()
+                return RunResult(
+                    status=Status.RUNTIME_ERROR, memory=0, time=0, return_code=0, outputs=None,
+                    errors=str(e),
+                )
+
+        self.db.commit()
+
+        # Execute the self.script as a single command and get the output
+        try:
+            print('Executing script:', self.script)
+            if self.script.strip().upper().startswith('SELECT'):
+                res = pd.read_sql_query(self.script, self.db).to_csv(index=False)
+            else:
+                cursor.executescript(self.script)
+                self.db.commit()
+                res = ''
+            print('Result:', res)
+        except (sqlite3.Error, pd.errors.ParserError, pd.errors.DatabaseError, ValueError) as e:
+            cursor.close()
+            return RunResult(
+                status=Status.RUNTIME_ERROR, memory=0, time=0, return_code=0, outputs=None,
+                errors=str(e),
+            )
+
+        # Read output files into the result
+        try:
+            r = RunResult(
+                status=Status.OK, memory=0, time=0, return_code=0, outputs=res,
+                output_files={
+                    filename: pd.read_sql_query(f'SELECT * FROM {filename}', self.db).to_csv(index=False)
+                    for filename in (test.target_files or {}).keys()
+                })
+            cursor.close()
+            return r
+        except (sqlite3.Error, pd.errors.ParserError, pd.errors.DatabaseError, ValueError) as e:
+            cursor.close()
+            return RunResult(
+                status=Status.RUNTIME_ERROR, memory=0, time=0, return_code=0, outputs=None,
+                errors=str(e),
+            )
+
+    def cleanup(self, test: TestCase) -> None:
+        """ Drops all the tables in the database """
+        cursor = self.db.cursor()
+        cursor.execute("SELECT name FROM sqlite_master WHERE type='table';")
+        tables = cursor.fetchall()
+
+        print(f'Dropping {len(tables)} tables')
+        for table_name in tables:
+            print('Dropping table:', table_name[0], end='...')
+            cursor.execute(f'DROP TABLE {table_name[0]}')
+            print('Done')
+
+        print('Saving changes')
+        self.db.commit()
+        cursor.close()

coderunners/lang/sqlite.Dockerfile

@@ -0,0 +1,18 @@
+FROM public.ecr.aws/lambda/python:3.12
+
+# Initial setup
+RUN pip install --upgrade pip
+RUN pip install awslambdaric -t "${LAMBDA_TASK_ROOT}"
+
+# Install dependencies
+COPY coderunners/requirements.txt ./
+RUN pip install -r requirements.txt -t "${LAMBDA_TASK_ROOT}"
+RUN python -m pip install --upgrade pandas
+
+# Setup source files
+COPY coderunners/*.py ${LAMBDA_TASK_ROOT}/coderunners/
+COPY models.py ${LAMBDA_TASK_ROOT}/
+
+# Run the lambda function handler
+ENTRYPOINT [ "python", "-m", "awslambdaric" ]
+CMD [ "coderunners.app.run_code_lambda" ]

coderunners/services.py

@@ -90,6 +90,7 @@ def check(self) -> SubmissionResult:
             self.test_cases[0],
             time_limit=self.time_limit, memory_limit_mb=self.memory_limit, output_limit_mb=self.output_limit
         )
+        executor.cleanup(self.test_cases[0])
         print('Done')
 
         # Process all tests
@@ -107,7 +108,7 @@ def check(self) -> SubmissionResult:
                 input_files=test.input_files, output_files=r.output_files, target_files=test.target_files,
                 input_assets=test.input_assets, output_assets=r.output_assets, target_assets=test.target_assets,
             ) if r.status == Status.OK else (r.status, 0, r.message)
-            print(f'Test {i} res: {r.status} => {r.score}')
+            print(f'Test {i} res: {r.status} => score {r.score}')
 
             # Clean up
             executor.cleanup(test)
@@ -135,8 +136,10 @@ def check(self) -> SubmissionResult:
                 test_results += [
                     RunResult(status=Status.SKIPPED, memory=0, time=0, return_code=0)
                 ] * (len(self.test_cases) - i - 1)
-                print('Expected:', test.target, test.target_files)
-                print('Actual:', test_results[-1].outputs, test_results[-1].output_files)
+                print('Expected:', test.target)
+                print('Actual:', test_results[-1].outputs)
+                print('Expected files:', test.target_files)
+                print('Actual files:', test_results[-1].output_files)
                 break
         print('test_results:', test_results)
         assert len(test_results) == len(self.test_cases)

template.yaml

@@ -514,6 +514,7 @@ Resources:
       DockerTag: js-v1
       DockerContext: ./
       Dockerfile: coderunners/lang/js.Dockerfile
+
   CodeRunnerJava:
     Type: AWS::Serverless::Function
     DependsOn: CodeRunnerMountTarget
@@ -534,6 +535,26 @@ Resources:
       DockerContext: ./
       Dockerfile: coderunners/lang/java.Dockerfile
 
+  CodeRunnerSQLite:
+    Type: AWS::Serverless::Function
+    DependsOn: CodeRunnerMountTarget
+    Properties:
+      FunctionName: CodeRunnerSQLite
+      PackageType: Image
+      Role: !GetAtt ContestantRole.Arn
+      VpcConfig:
+        SecurityGroupIds:
+          - !GetAtt JudgeVPC.DefaultSecurityGroup
+        SubnetIds:
+          - !Ref CodeRunnerPrivateSubnet
+      FileSystemConfigs:
+        - Arn: !GetAtt AccessPointResource.Arn
+          LocalMountPath: '/mnt/efs'
+    Metadata:
+      DockerTag: sqlite-v1
+      DockerContext: ./
+      Dockerfile: coderunners/lang/sqlite.Dockerfile
+
 Outputs:
   # ServerlessRestApi is an implicit API created out of Events key under Serverless::Function
   # Find out more about other implicit resources you can reference within SAM
@@ -569,6 +590,9 @@ Outputs:
   CodeRunnerJava:
     Description: 'AWS Lambda for executing a Java code and getting the outputs'
     Value: !GetAtt CodeRunnerJava.Arn
+  CodeRunnerSQLite:
+    Description: 'AWS Lambda for executing a SQL/SQLite code and getting the results with tables'
+    Value: !GetAtt CodeRunnerSQLite.Arn
 
   SyncS3WithEFSName:
     Description: 'AWS Lambda for syncing S3 bucket for test cases with EFS'