feat: Link Datadoc to GitHub

package.json

@@ -1,6 +1,6 @@
 {
     "name": "querybook",
-    "version": "3.34.2",
+    "version": "3.35.0",
     "description": "A Big Data Webapp",
     "private": true,
     "scripts": {

querybook/migrations/versions/a8bb37120302_add_github_link.py

@@ -0,0 +1,52 @@
+"""Add Github Link
+
+Revision ID: a8bb37120302
+Revises: f7b11b3e3a95
+Create Date: 2024-10-18 15:50:51.471760
+
+"""
+
+from alembic import op
+import sqlalchemy as sa
+
+# revision identifiers, used by Alembic.
+revision = "a8bb37120302"
+down_revision = "f7b11b3e3a95"
+branch_labels = None
+depends_on = None
+
+
+def upgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.create_table(
+        "github_link",
+        sa.Column("id", sa.Integer(), autoincrement=True, nullable=False),
+        sa.Column("datadoc_id", sa.Integer(), nullable=False),
+        sa.Column("user_id", sa.Integer(), nullable=False),
+        sa.Column("repo_url", sa.String(length=255), nullable=False),
+        sa.Column("branch", sa.String(length=255), nullable=False),
+        sa.Column("file_path", sa.String(length=255), nullable=False),
+        sa.Column(
+            "created_at", sa.DateTime(), server_default=sa.text("now()"), nullable=False
+        ),
+        sa.Column(
+            "updated_at", sa.DateTime(), server_default=sa.text("now()"), nullable=False
+        ),
+        sa.ForeignKeyConstraint(
+            ["datadoc_id"],
+            ["data_doc.id"],
+        ),
+        sa.ForeignKeyConstraint(
+            ["user_id"],
+            ["user.id"],
+        ),
+        sa.PrimaryKeyConstraint("id"),
+        sa.UniqueConstraint("datadoc_id"),
+    )
+    # ### end Alembic commands ###
+
+
+def downgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_table("github_link")
+    # ### end Alembic commands ###

querybook/server/datasources/github.py

@@ -1,6 +1,8 @@
 from app.datasource import register
 from lib.github_integration.github_integration import get_github_manager
 from typing import Dict
+from logic import github as logic
+from flask_login import current_user
 
 
 @register("/github/auth/", methods=["GET"])
@@ -14,3 +16,19 @@ def is_github_authenticated() -> str:
     github_manager = get_github_manager()
     is_authenticated = github_manager.get_github_token() is not None
     return {"is_authenticated": is_authenticated}
+
+
+@register("/github/datadocs/<int:datadoc_id>/link/", methods=["POST"])
+def link_datadoc_to_github(
+    datadoc_id: int,
+    repo_url: str,
+    branch: str,
+    file_path: str,
+) -> Dict:
+    return logic.create_repo_link(
+        datadoc_id=datadoc_id,
+        user_id=current_user.id,
+        repo_url=repo_url,
+        branch=branch,
+        file_path=file_path,
+    )

querybook/server/logic/github.py

@@ -0,0 +1,44 @@
+from app.db import with_session
+from models.github import GitHubLink
+from models.datadoc import DataDoc
+
+
+@with_session
+def create_repo_link(
+    datadoc_id: int,
+    user_id: int,
+    repo_url: str,
+    branch: str,
+    file_path: str,
+    commit=True,
+    session=None,
+):
+    datadoc = DataDoc.get(id=datadoc_id, session=session)
+    assert datadoc is not None, f"DataDoc with id {datadoc_id} not found"
+
+    github_link = GitHubLink.get(datadoc_id=datadoc_id, session=session)
+    assert (
+        github_link is None
+    ), f"GitHub link for DataDoc with id {datadoc_id} already exists"
+
+    github_link = GitHubLink.create(
+        {
+            "datadoc_id": datadoc_id,
+            "user_id": user_id,
+            "repo_url": repo_url,
+            "branch": branch,
+            "file_path": file_path,
+        },
+        commit=commit,
+        session=session,
+    )
+    return github_link
+
+
+@with_session
+def get_repo_link(datadoc_id: int, session=None):
+    github_link = GitHubLink.get(datadoc_id=datadoc_id, session=session)
+    assert (
+        github_link is not None
+    ), f"GitHub link for DataDoc with id {datadoc_id} not found"
+    return github_link

querybook/server/models/__init__.py

@@ -15,3 +15,4 @@
 from .data_element import *
 from .comment import *
 from .survey import *
+from .github import *

querybook/server/models/github.py

@@ -0,0 +1,41 @@
+import sqlalchemy as sql
+from sqlalchemy.sql import func
+from lib.sqlalchemy import CRUDMixin
+from sqlalchemy.orm import backref, relationship
+from app import db
+
+Base = db.Base
+
+
+class GitHubLink(Base, CRUDMixin):
+    __tablename__ = "github_link"
+    id = sql.Column(sql.Integer, primary_key=True, autoincrement=True)
+    datadoc_id = sql.Column(
+        sql.Integer, sql.ForeignKey("data_doc.id"), nullable=False, unique=True
+    )
+    user_id = sql.Column(sql.Integer, sql.ForeignKey("user.id"), nullable=False)
+    repo_url = sql.Column(sql.String(255), nullable=False)
+    branch = sql.Column(sql.String(255), nullable=False)
+    file_path = sql.Column(sql.String(255), nullable=False)
+    created_at = sql.Column(sql.DateTime, server_default=func.now(), nullable=False)
+    updated_at = sql.Column(
+        sql.DateTime, server_default=func.now(), onupdate=func.now(), nullable=False
+    )
+
+    datadoc = relationship(
+        "DataDoc",
+        backref=backref("github_link", uselist=False, cascade="all, delete-orphan"),
+    )
+    user = relationship("User", backref=backref("github_link", uselist=False))
+
+    def to_dict(self):
+        return {
+            "id": self.id,
+            "datadoc_id": self.datadoc_id,
+            "user_id": self.user_id,
+            "repo_url": self.repo_url,
+            "branch": self.branch,
+            "file_path": self.file_path,
+            "created_at": self.created_at,
+            "updated_at": self.updated_at,
+        }