diff --git a/schema/versions/44d5bbbf2615_workspaces_init.py b/schema/versions/44d5bbbf2615_workspaces_init.py
index c2a58a44c81c6220fbbfb45bcf48cdc1bffd4cef..0b3922aa3d64c1ac96c2eae9e82eef3c62c7cbea 100644
--- a/schema/versions/44d5bbbf2615_workspaces_init.py
+++ b/schema/versions/44d5bbbf2615_workspaces_init.py
@@ -16,20 +16,22 @@ depends_on = None
def upgrade():
op.create_table('workflows',
- sa.Column('workflow_id', sa.Integer, primary_key=True, autoincrement=True,
- comment='the unique id of the workflow. auto-generated.'),
- sa.Column('name', sa.String, unique=True, comment='a human-readable name for the workflow.'),
+ sa.Column('workflow_name', sa.String, primary_key=True, comment='the human-readable name for the workflow.'),
comment='A listing of the available workflows in the system.')
+ op.create_table('workflow_templates',
+ sa.Column('filename', sa.String, primary_key=True, comment='the filename of the template'),
+ sa.Column('content', sa.String, nullable=False, comment='the content of the template'),
+ sa.Column('workflow_name', sa.String, sa.ForeignKey('workflows.workflow_name'), primary_key=True,
+ comment='the name of the workflow this template belongs to'),
+ comment='Templates associated with workflows')
+
op.create_table('workflow_requests',
sa.Column('workflow_request_id', sa.Integer, primary_key=True, autoincrement=True,
comment='the unique id of the request. auto-generated'),
- sa.Column('job_id', sa.Integer,
- comment='the id of the job that this request generted in the HTCondor system.'),
- sa.Column('workflow_id', sa.Integer, sa.ForeignKey('workflows.workflow_id'),
- comment='the id of the workflow used in this request.'),
+ sa.Column('workflow_name', sa.String, sa.ForeignKey('workflows.workflow_name')),
sa.Column('argument', sa.JSON, comment='the argument(s) used for the workflow in this request.'),
- comment='A listing of requests for workflows and te resulting job ids.')
+ comment='A listing of requests for workflow execution.')
op.create_table('workflow_request_files',
sa.Column('workflow_request_id', sa.Integer, sa.ForeignKey('workflow_requests.workflow_request_id'),
@@ -37,11 +39,17 @@ def upgrade():
sa.Column('filename', sa.String, comment='the name of this file', nullable=False),
sa.Column('content', sa.String, comment='the contents of the file', nullable=False),
comment='A man-to-many mapping table tracking which files were used for workflow requests.')
+
op.create_unique_constraint('workflow_request_filenames_uniq', 'workflow_request_files',
['workflow_request_id', 'filename'])
+ # populate with some initial data
+ op.execute("INSERT INTO workflows (workflow_name) VALUES ('null')")
+ op.execute("INSERT INTO workflow_templates (workflow_name, filename, content) VALUES ('null', 'null.sh', 'null {{arguments}}')")
+
def downgrade():
op.drop_table('workflow_request_files')
op.drop_table('workflow_requests')
+ op.drop_table('workflow_templates')
op.drop_table('workflows')
\ No newline at end of file
diff --git a/services/workflow/src/workflow/server.py b/services/workflow/src/workflow/server.py
index b5f8d0495e0ed6d15ff1ce869311cdaef0d944c5..f4c8c0147d0ecae33fc3cfb0abaf5cd2ebbc8b22 100644
--- a/services/workflow/src/workflow/server.py
+++ b/services/workflow/src/workflow/server.py
@@ -1,4 +1,10 @@
-import sqlalchemy
+import inspect
+from abc import ABC, abstractmethod
+from dataclasses import dataclass
+from pathlib import Path
+from typing import List, Dict
+
+import sqlalchemy as sa
import zope.sqlalchemy
from pycapo import CapoConfig
from pyramid.view import view_config, view_defaults
@@ -7,20 +13,19 @@ from pyramid_beaker import session_factory_from_settings
from pyramid.renderers import JSONP
from sqlalchemy import create_engine
from sqlalchemy.ext.declarative import declarative_base
-from sqlalchemy.orm import sessionmaker
-
+from sqlalchemy.orm import sessionmaker, relationship
"""
Work done:
☑ Initial sketch of using SQL Alchemy with Pyramid
☑ Sketch of first routes we need
+ ☑ Bring over interfaces from wksp0 project
+ ☑ Need to flesh out the object model—requests have their own files, workflows have templates
+ ☑ Separate REST hierarchy for workflow definitions
Work to do:
- ☠Bring over interfaces from wksp0 project
- ☠Need to flesh out the object model—requests have their own files, workflows have templates
- ☠Separate REST hierarchy for workflow definitions
☠Actually do the work, preferably in model classes, not REST API directly
☠Separate this into separate modules, once it makes sense to people how it works
☠Workflow initiation CLI
@@ -31,6 +36,52 @@ To consider:
"""
+# ---------------------------------------------------------
+#
+# I N T E R F A C E S
+#
+# ---------------------------------------------------------
+
+
+class WorkflowService(ABC):
+ """
+ Executes workflows; should be a freestanding service.
+ """
+ @abstractmethod
+ def execute(self, workflow_name: str, argument: Dict, files: List[Path]):
+ """
+ Execute this workflow against these files.
+
+ :param workflow_name: name of the workflow to run
+ :param argument: extra argument (a JSON object)
+ :param files: some extra files the workflow should consider
+ :return: a stream of events from this workflow
+ """
+ raise NotImplementedError(f'{self.__class__.__name__}.{inspect.stack()[0][3]}')
+
+
+class WorkflowInfo(ABC):
+ """
+ Holds information about workflows.
+ """
+ @abstractmethod
+ def lookup_workflow_definition(self, name: str) -> 'Workflow':
+ """
+ Look up the workflow with this name.
+
+ :param name: Workflow name
+ :return: Workflow instance
+ """
+ raise NotImplementedError(f'{self.__class__.__name__}.{inspect.stack()[0][3]}')
+
+
+class AbstractFile:
+ def __init__(self, filename, content):
+ self.filename, self.content = filename, content
+
+ def write_to(self, directory: Path):
+ raise NotImplementedError('Actually save the file into the directory passed')
+
# ---------------------------------------------------------
#
@@ -38,24 +89,87 @@ To consider:
#
# ---------------------------------------------------------
+
Base = declarative_base()
class Workflow(Base):
"""
- A Workflow is a suite of tasks defined by templates that must be executed together with some user-submitted files.
+ A Workflow is a suite of tasks defined by templates that must be
+ executed together with some user-submitted files.
"""
__tablename__ = 'workflows'
- id = sqlalchemy.Column('workflow_id', sqlalchemy.String, primary_key=True, autoincrement=True)
- name = sqlalchemy.Column('name', sqlalchemy.String)
+ workflow_name = sa.Column('workflow_name', sa.String, primary_key=True)
+ templates = relationship('WorkflowTemplate', backref='workflow')
+ requests = relationship('WorkflowRequest', backref='workflow')
def __json__(self, request):
- return {'id': self.id, 'name': self.name}
+ return {'workflow_name': self.workflow_name}
+
+ def render_templates(self, argument: Dict, files: List[Path]) -> List[AbstractFile]:
+ """
+ Render the templates associated with this workflow
+ :param argument: the workflow argument JSON
+ :param files: the files to be processed
+ :return: a list of rendered templates
+ """
+ raise NotImplementedError(f'{self.__class__.__name__}.{inspect.stack()[0][3]}')
+
+
+class WorkflowTemplate(Base):
+ """
+ Workflow Templates are Mustache-formatted files associated with a given Workflow
+ """
+ __tablename__ = 'workflow_templates'
+ filename = sa.Column('filename', sa.String, primary_key=True)
+ content = sa.Column('content', sa.String, nullable=False)
+ workflow_name = sa.Column('workflow_name', sa.String, sa.ForeignKey('workflows.workflow_name'), primary_key=True)
+
+ def render(self, argument: Dict) -> AbstractFile:
+ raise NotImplementedError('Add mustache to render this template')
+
+
+class WorkflowRequest(Base):
+ """
+ Workflow Requests invoke a particular workflow with a particular argument and some user-supplied files.
+ """
+ __tablename__ = 'workflow_requests'
+ workflow_request_id = sa.Column('workflow_request_id', sa.Integer, primary_key=True, autoincrement=True)
+ workflow_name = sa.Column('workflow_name', sa.String, sa.ForeignKey('workflows.workflow_name'))
+ argument = sa.Column('argument', sa.JSON)
+ files = relationship('WorkflowRequestFile', backref='request')
+
+
+class WorkflowRequestFile(Base):
+ __tablename__ = 'workflow_request_files'
+ workflow_request_id = sa.Column('workflow_request_id', sa.Integer, sa.ForeignKey('workflow_requests.workflow_request_id'),
+ primary_key=True)
+ filename = sa.Column('filename', sa.String, primary_key=True)
+ content = sa.Column('content', sa.String, nullable=False)
+
+ @property
+ def file(self) -> AbstractFile:
+ return AbstractFile(self.filename, self.content)
WORKFLOWS = [{'id': 1, 'name': 'foo', 'files': {'file.txt': {'id': 1, 'name': 'file.txt', 'content': 'Hello, world!'}}}]
+class DBWorkflowInfo(WorkflowInfo):
+ """
+ Implements WorkflowInfo, backed by a relational database, using SQL Alchemy.
+ """
+
+ def __init__(self, session):
+ self.session = session
+
+ def lookup_workflow_definition(self, name: str) -> Workflow:
+ return self.session.query(Workflow).get(name)
+
+ def all_workflows(self) -> List[Workflow]:
+ return self.session.query(Workflow).all()
+
+
# ---------------------------------------------------------
#
# S E R V I C E S
@@ -65,7 +179,7 @@ WORKFLOWS = [{'id': 1, 'name': 'foo', 'files': {'file.txt': {'id': 1, 'name': 'f
def lookup_workflow(request):
# this should change to use SQL Alchemy
- return WORKFLOWS[int(request.matchdict['id'])]
+ return request.info.lookup_workflow_definition(request.matchdict['id'])
def lookup_file(request):
@@ -91,7 +205,7 @@ class WorkflowService:
Audience: front-end
:return:
"""
- return self.request.dbsession.query(Workflow).all()
+ return self.request.info.all_workflows()
@view_config(request_method='POST')
def create_workflow(self):
@@ -220,7 +334,6 @@ def get_tm_session(session_factory, transaction_manager):
# ---------------------------------------------------------
-
def main(global_config, **settings):
with Configurator(settings=settings) as config:
session_factory = session_factory_from_settings(settings)
@@ -241,8 +354,8 @@ def main(global_config, **settings):
# make request.dbsession available for use in Pyramid
config.add_request_method(
# r.tm is the transaction manager used by pyramid_tm
- lambda r: get_tm_session(session_factory, r.tm),
- 'dbsession',
+ lambda request: DBWorkflowInfo(get_tm_session(session_factory, request.tm)),
+ 'info',
reify=True
)