Merge pull request #7 from zhangvi7/github/datadoc-link

feat: Link Datadoc to GitHub Directory

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/aa328ae9dced_add_github_datadoc_link.py

@@ -0,0 +1,56 @@
+"""Add GitHub Datadoc Link
+
+Revision ID: aa328ae9dced
+Revises: f7b11b3e3a95
+Create Date: 2024-10-23 21:04:55.052696
+
+"""
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision = "aa328ae9dced"
+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(
+            "directory",
+            sa.String(length=255),
+            nullable=False,
+            server_default="datadocs",
+        ),
+        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.github import github_manager
 from typing import Dict
+from logic import github as logic
+from flask_login import current_user
 
 
 @register("/github/auth/", methods=["GET"])
@@ -12,3 +14,13 @@ def connect_github() -> Dict[str, str]:
 def is_github_authenticated() -> str:
     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,
+    directory: str,
+) -> Dict:
+    return logic.create_repo_link(
+        datadoc_id=datadoc_id, user_id=current_user.id, directory=directory
+    )

querybook/server/lib/github/serializers.py

@@ -0,0 +1,145 @@
+import yaml
+import re
+from typing import List
+from models.datadoc import DataDoc, DataCell
+from const.data_doc import DataCellType
+from datetime import datetime, timezone
+
+
+def parse_datetime_as_utc(date_str: str) -> datetime:
+    """
+    Parse the given date string to a datetime object in UTC.
+    """
+    if isinstance(date_str, datetime):
+        return date_str.astimezone(timezone.utc)
+    if date_str:
+        return datetime.fromisoformat(date_str).astimezone(timezone.utc)
+    return datetime.now(timezone.utc).replace(tzinfo=timezone.utc)
+
+
+def serialize_datadoc_to_markdown(datadoc: DataDoc) -> str:
+    # Serialize DataDoc metadata to YAML front matter for readability
+    datadoc_metadata = {
+        "id": datadoc.id,
+        "environment_id": datadoc.environment_id,
+        "public": datadoc.public,
+        "archived": datadoc.archived,
+        "owner_uid": datadoc.owner_uid,
+        "created_at": datadoc.created_at.isoformat() if datadoc.created_at else None,
+        "updated_at": datadoc.updated_at.isoformat() if datadoc.updated_at else None,
+        "meta": datadoc.meta,
+        "title": datadoc.title,
+    }
+    try:
+        front_matter = (
+            f"---\n{yaml.dump(datadoc_metadata, default_flow_style=False)}---\n\n"
+        )
+    except yaml.YAMLError as e:
+        raise ValueError(f"Error serializing DataDoc metadata to YAML: {e}")
+
+    title = f"# {datadoc.title}\n\n"
+    content = serialize_datacells(datadoc.cells)
+    markdown_content = front_matter + title + content
+    return markdown_content
+
+
+def serialize_datacells(cells: List[DataCell]) -> str:
+    lines = []
+    for cell in cells:
+        # Since GitHub's Markdown renderer does not recognize multiple --- blocks as separate YAML sections,
+        # we serialize cell metadata in HTML comment to hide it from rendered view
+        cell_metadata = {
+            "id": cell.id,
+            "cell_type": cell.cell_type.name.lower(),
+            "created_at": cell.created_at.isoformat() if cell.created_at else None,
+            "updated_at": cell.updated_at.isoformat() if cell.updated_at else None,
+            "meta": cell.meta,
+        }
+        try:
+            cell_metadata_yaml = yaml.dump(cell_metadata, default_flow_style=False)
+        except yaml.YAMLError as e:
+            raise ValueError(f"Error serializing cell metadata to YAML: {e}")
+
+        cell_metadata_comment = f"<!--\n{cell_metadata_yaml.strip()}\n-->\n"
+
+        cell_content = serialize_cell_content(cell)
+        lines.append(cell_metadata_comment + cell_content)
+
+    return "\n\n".join(lines)
+
+
+def serialize_cell_content(cell: DataCell) -> str:
+    cell_meta = cell.meta or {}
+    if cell.cell_type == DataCellType.query:
+        query_title = cell_meta.get("title", "Query")
+        return f"## Query: {query_title}\n\n```sql\n{cell.context.strip()}\n```\n"
+    elif cell.cell_type == DataCellType.text:
+        return f"## Text\n\n```text\n{cell.context.strip()}\n```\n"
+    elif cell.cell_type == DataCellType.chart:
+        return "## Chart\n\n```text\n*Chart generated from the metadata.*\n```\n"
+
+
+def deserialize_datadoc_from_markdown(markdown_str: str) -> DataDoc:
+    front_matter, content = extract_front_matter(markdown_str)
+    datadoc = create_datadoc_from_metadata(front_matter)
+    datadoc.cells = deserialize_datadoc_content(content)
+    return datadoc
+
+
+def extract_front_matter(markdown_str: str):
+    front_matter_pattern = re.compile(r"^---\n(.*?)\n---\n\n", re.DOTALL)
+    match = front_matter_pattern.match(markdown_str)
+    if match:
+        front_matter_str = match.group(1)
+        content = markdown_str[match.end() :]
+        try:
+            front_matter = yaml.safe_load(front_matter_str)
+        except yaml.YAMLError as e:
+            raise ValueError(f"Error parsing front matter YAML: {e}")
+    else:
+        raise ValueError("Invalid Markdown format: Missing front matter.")
+    return front_matter, content
+
+
+def create_datadoc_from_metadata(metadata: dict) -> DataDoc:
+    datadoc = DataDoc(
+        id=metadata.get("id"),
+        environment_id=metadata.get("environment_id"),
+        public=metadata.get("public", True),
+        archived=metadata.get("archived", False),
+        owner_uid=metadata.get("owner_uid"),
+        created_at=parse_datetime_as_utc(metadata.get("created_at")),
+        updated_at=parse_datetime_as_utc(metadata.get("updated_at")),
+        title=metadata.get("title", ""),
+    )
+    datadoc.meta = metadata.get("meta", {})
+    return datadoc
+
+
+def deserialize_datadoc_content(content_str: str) -> List[DataCell]:
+    cells = []
+    # Pattern to match cell metadata in HTML comments and the following content
+    pattern = re.compile(r"<!--\n(.*?)\n-->\n## .*?\n\n```.*?\n(.*?)\n```", re.DOTALL)
+    matches = pattern.finditer(content_str)
+    for match in matches:
+        metadata_str = match.group(1)
+        cell_content = match.group(2)
+        try:
+            metadata = yaml.safe_load(metadata_str)
+        except yaml.YAMLError as e:
+            raise ValueError(f"Error parsing cell metadata YAML: {e}")
+
+        cell_type_str = metadata.get("cell_type", "query").lower()
+        cell_type = DataCellType[cell_type_str]
+        cell = DataCell(
+            id=metadata.get("id"),
+            cell_type=cell_type,
+            context=(
+                cell_content.strip() if cell_type != DataCellType.chart else None
+            ),  # Charts are generated from the metadata, and not from content
+            created_at=parse_datetime_as_utc(metadata.get("created_at")),
+            updated_at=parse_datetime_as_utc(metadata.get("updated_at")),
+            meta=metadata.get("meta", {}),
+        )
+        cells.append(cell)
+    return cells

querybook/server/logic/github.py

@@ -0,0 +1,40 @@
+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,
+    directory: 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,
+            "directory": directory,
+        },
+        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 +1,37 @@
+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)
+    directory = sql.Column(sql.String(255), nullable=False, default="datadocs")
+    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,
+            "directory": self.directory,
+            "created_at": self.created_at,
+            "updated_at": self.updated_at,
+        }