Fix more bugs in publishing

package.json

@@ -8,7 +8,7 @@
   "scripts": {
     "pre-commit-version": "pnpm changeset",
     "version": "./scripts/changeset-version.py version",
-    "publish": "./scripts/changeset-version.py publish --no-js --tag"
+    "publish": "./scripts/changeset-version.py publish"
   },
   "devDependencies": {
     "prettier": "^3.6.2",

pnpm-lock.yaml

@@ -21,21 +21,7 @@ importers:
         specifier: ^3.6.2
         version: 3.6.2
 
-  py:
-    devDependencies:
-      changesets:
-        specifier: ^1.0.2
-        version: 1.0.2
-
-  py/llama_parse:
-    dependencies:
-      llama-cloud-services-py:
-        specifier: workspace:*
-        version: link:..
-    devDependencies:
-      changesets:
-        specifier: ^1.0.2
-        version: 1.0.2
+  py: {}
 
   ts/e2e-tests:
     devDependencies:

pnpm-workspace.yaml

@@ -1,4 +1,3 @@
 packages:
   - "ts/*"
   - "py"
-  - "py/*"

py/CHANGELOG.md

@@ -1,17 +1,5 @@
 # llama-cloud-services-py
 
-## 0.6.72
-
-### Patch Changes
-
-- ad6734b: Fixup and test versioning
-
-## 0.6.71
-
-### Patch Changes
-
-- 51011b9: Escape dollar signs in jupyter notebooks
-
 ## 0.6.70
 
 ### Patch Changes

py/llama_cloud_services/parse/types.py

@@ -4,10 +4,7 @@ import re
 from pydantic import BaseModel, Field, SerializeAsAny
 from typing import Dict, Any, List, Optional
 
-from llama_cloud_services.parse.utils import (
-    make_api_request,
-    is_jupyter,
-)
+from llama_cloud_services.parse.utils import make_api_request
 from llama_index.core.async_utils import asyncio_run
 from llama_index.core.schema import Document, ImageDocument, ImageNode, TextNode
 
@@ -261,24 +258,6 @@ class JobResult(BaseModel):
         documents = await self.aget_text_documents(split_by_page)
         return [TextNode(text=doc.text, metadata=doc.metadata) for doc in documents]
 
-    def _format_markdown_for_notebook(self, text: Optional[str]) -> Optional[str]:
-        """Format markdown text for Jupyter notebook display by escaping dollar signs."""
-        if text is None:
-            return None
-
-        def escape_dollar_signs(text: str) -> str:
-            """Escape dollar signs in text to prevent Jupyter from interpreting them as LaTeX.
-
-            Args:
-                text: The text to escape
-
-            Returns:
-                Text with dollar signs escaped
-            """
-            return text.replace("$", r"\$")
-
-        return escape_dollar_signs(text)
-
     def get_markdown_documents(self, split_by_page: bool = False) -> List[Document]:
         """
         Get the markdown documents from the job.
@@ -289,22 +268,17 @@ class JobResult(BaseModel):
         if split_by_page:
             return [
                 Document(
-                    text=self._format_markdown_for_notebook(page.md)
-                    if is_jupyter()
-                    else page.md,
+                    text=page.md,
                     metadata={"page_number": page.page, "file_name": self.file_name},
                 )
                 for page in self.pages
             ]
         else:
-            text = self._page_separator.join(
-                [page.md if page.md is not None else "" for page in self.pages]
-            )
             return [
                 Document(
-                    text=self._format_markdown_for_notebook(text)
-                    if is_jupyter()
-                    else text,
+                    text=self._page_separator.join(
+                        [page.md if page.md is not None else "" for page in self.pages]
+                    ),
                     metadata={"file_name": self.file_name},
                 )
             ]
@@ -354,10 +328,7 @@ class JobResult(BaseModel):
         """
         url = f"{self._base_url}/api/v1/parsing/job/{self.job_id}/result/raw/markdown"
         response = await make_api_request(self._client, "GET", url)
-        markdown = response.content.decode("utf-8")
-        return (
-            self._format_markdown_for_notebook(markdown) if is_jupyter() else markdown
-        )
+        return response.content.decode("utf-8")
 
     def get_text(self) -> str:
         """

py/llama_cloud_services/parse/utils.py

@@ -1,4 +1,3 @@
-import functools
 import httpx
 import itertools
 import logging
@@ -357,17 +356,6 @@ def partition_pages(
             return
 
 
-@functools.lru_cache(maxsize=1)
-def is_jupyter() -> bool:
-    """Check if we're running in a Jupyter environment."""
-    try:
-        from IPython import get_ipython
-
-        return get_ipython().__class__.__name__ == "ZMQInteractiveShell"
-    except (ImportError, AttributeError):
-        return False
-
-
 def extract_tables_from_json_results(
     json_results: List[dict], download_path: str
 ) -> List[str]:

py/llama_parse/CHANGELOG.md

@@ -1,8 +0,0 @@
-# llama_parse
-
-## 0.6.72
-
-### Patch Changes
-
-- Updated dependencies [ad6734b]
-  - llama-cloud-services-py@0.6.72

py/llama_parse/package.json

@@ -1,20 +0,0 @@
-{
-  "name": "llama_parse",
-  "version": "0.6.72",
-  "description": "",
-  "main": "index.js",
-  "private": false,
-  "scripts": {
-    "test": "echo \"Error: no test specified\" && exit 1"
-  },
-  "dependencies": {
-    "llama-cloud-services-py": "workspace:*"
-  },
-  "keywords": [],
-  "author": "",
-  "license": "ISC",
-  "packageManager": "pnpm@10.11.1",
-  "devDependencies": {
-    "changesets": "^1.0.2"
-  }
-}

py/llama_parse/pyproject.toml

@@ -11,13 +11,13 @@ dev = [
 
 [project]
 name = "llama-parse"
-version = "0.6.72"
+version = "0.6.70"
 description = "Parse files into RAG-Optimized formats."
 authors = [{name = "Logan Markewich", email = "logan@llamaindex.ai"}]
 requires-python = ">=3.9,<4.0"
 readme = "README.md"
 license = "MIT"
-dependencies = ["llama-cloud-services>=0.6.72"]
+dependencies = ["llama-cloud-services>=0.6.70"]
 
 [project.scripts]
 llama-parse = "llama_parse.cli.main:parse"

py/package.json

@@ -1,10 +1,7 @@
 {
   "name": "llama-cloud-services-py",
-  "version": "0.6.72",
-  "private": false,
+  "version": "0.6.70",
+  "private": "true",
   "license": "MIT",
-  "scripts": {},
-  "devDependencies": {
-    "changesets": "^1.0.2"
-  }
+  "scripts": {}
 }

py/pyproject.toml

@@ -19,7 +19,7 @@ dev = [
 
 [project]
 name = "llama-cloud-services"
-version = "0.6.72"
+version = "0.6.70"
 description = "Tailored SDK clients for LlamaCloud services."
 authors = [{name = "Logan Markewich", email = "logan@runllama.ai"}]
 requires-python = ">=3.9,<4.0"

py/uv.lock

@@ -1,5 +1,5 @@
 version = 1
-revision = 3
+revision = 2
 requires-python = ">=3.9, <4.0"
 resolution-markers = [
     "python_full_version >= '3.14'",
@@ -1596,7 +1596,7 @@ wheels = [
 
 [[package]]
 name = "llama-cloud-services"
-version = "0.6.70"
+version = "0.6.69"
 source = { editable = "." }
 dependencies = [
     { name = "click", version = "8.1.8", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.10'" },

scripts/changeset-version.py

@@ -12,15 +12,14 @@ There's 2 things this does:
 
 """
 
-from dataclasses import dataclass
 import json
 import os
 import subprocess
+import sys
 from pathlib import Path
-from typing import Any, List, cast
+from typing import List
 import urllib.request
 import urllib.error
-import re
 
 import click
 import tomlkit
@@ -34,97 +33,40 @@ def _run_command(
     subprocess.run(cmd, check=True, text=True, cwd=cwd or Path.cwd(), env=env)
 
 
-def _run_and_capture(
-    cmd: List[str], cwd: Path | None = None, env: dict[str, str] | None = None
-) -> str:
-    """Run a command and return stdout as text, raising on failure."""
-    result = subprocess.run(
-        cmd,
-        check=True,
-        text=True,
-        cwd=cwd or Path.cwd(),
-        env=env,
-        capture_output=True,
-    )
-    return result.stdout
+def update_python_versions(version: str) -> None:
+    """llama-cloud-services and llama-parse share a version. llama-parse is just a silly sidecar that proxies to llama-cloud-services
+    for compatibility.
 
-
-@dataclass
-class Package:
-    name: str
-    version: str
-    path: Path
-
-    def python_package_name(self) -> str | None:
-        if "/py/" in str(self.path) or str(self.path).endswith("/py"):
-            return self.name.removesuffix("-py")
-        return None
-
-
-def _get_pnpm_workspace_packages() -> list[Package]:
-    """Return directories for all workspace packages from pnpm list JSON output."""
-    output = _run_and_capture(["pnpm", "list", "-r", "--depth=-1", "--json"])
-
-    data = cast(list[dict[str, Any]], json.loads(output))
-    packages: list[Package] = [
-        Package(name=data["name"], version=data["version"], path=Path(data["path"]))
-        for data in data
-    ]
-    return packages
-
-
-def _sync_package_version_with_pyproject(
-    package_dir: Path, packages: dict[str, Package], js_package_name: str
-) -> None:
-    """Sync version from package.json to pyproject.toml.
-
-    Returns True if pyproject was changed, else False.
+    This function updates the version in both pyproject.toml files.
     """
-    pyproject_path = package_dir / "pyproject.toml"
-    if not pyproject_path.exists():
-        return
+    # Update main pyproject.toml
+    main_path = Path("py/pyproject.toml")
+    main_content = main_path.read_text()
+    main_doc = tomlkit.parse(main_content)
+    if main_doc["project"]["version"] != version:
+        click.echo(f"Updating llama-cloud-services version to {version}")
+        main_doc["project"]["version"] = version
+    main_path.write_text(tomlkit.dumps(main_doc))
 
-    package_version = packages[js_package_name].version
-    py_doc = tomlkit.parse(pyproject_path.read_text())
+    # Update llama_parse/pyproject.toml
+    parse_path = Path("py/llama_parse/pyproject.toml")
+    parse_content = parse_path.read_text()
+    parse_doc = tomlkit.parse(parse_content)
+    if parse_doc["project"]["version"] != version:
+        click.echo(f"Updating llama-parse version to {version}")
+        parse_doc["project"]["version"] = version
+        parse_path.write_text(tomlkit.dumps(parse_doc))
 
-    by_python_name = {
-        pkg.python_package_name(): pkg
-        for pkg in packages.values()
-        if pkg.python_package_name()
-    }
+    # Update the dependency reference
+    dependencies = parse_doc["project"]["dependencies"]
+    for i, dep in enumerate(dependencies):
+        if isinstance(dep, str) and dep.startswith("llama-cloud-services"):
+            dependencies[i] = f"llama-cloud-services>={version}"
+            break
 
-    current_version = py_doc["project"]["version"]
-    assert isinstance(current_version, str)
+    parse_path.write_text(tomlkit.dumps(parse_doc))
 
-    # update workspace dependency strings by replacing the first version after == or >=
-    deps = py_doc["project"]["dependencies"] or []
-    changed = False
-    for i, dep in enumerate(deps):
-        if not isinstance(dep, str):
-            continue
-        pkg = (cast(str, dep).split("==")[0]).split(">=")[0]
-        if pkg not in by_python_name:
-            continue
-        target_version = by_python_name[pkg].version
-        new_dep = re.sub(
-            r"(==|>=)\s*([0-9A-Za-z_.+-]+)",
-            lambda m: m.group(1) + target_version,
-            dep,
-            count=1,
-        )
-        if new_dep != dep:
-            deps[i] = new_dep
-            changed = True
-
-    if current_version != package_version:
-        py_doc["project"]["version"] = package_version
-        changed = True
-
-    if changed:
-        pyproject_path.write_text(tomlkit.dumps(py_doc))
-        click.echo(
-            f"Updated {pyproject_path} version to {package_version} and synced dependency specs"
-        )
+    click.echo(f"Updated Python packages to version {version}")
 
 
 def lock_python_dependencies() -> None:
@@ -144,64 +86,57 @@ def cli() -> None:
 
 @cli.command()
 def version() -> None:
-    """Apply changeset versions, then sync versions for co-located JS/Py packages.
-
-    - Runs changesets to bump package.json versions.
-    - Discovers all workspace packages via pnpm.
-    - For any directory containing both package.json and pyproject.toml, and with
-      package.json private: false, set pyproject [project].version to match the JS version.
-    - If a pyproject is updated, run `uv sync` in that directory to update its lock file.
-    """
-    # Ensure we're at the repo root
-    os.chdir(Path(__file__).parent.parent)
-
-    # First, run changeset version to update all package.json files
+    """Apply changeset versions, and propagate them to Python packages."""
+    # First, run changeset version to update all package.json files (including py/package.json)
     _run_command(["npx", "@changesets/cli", "version"])
 
-    # Enumerate workspace packages and perform syncs
-    packages = _get_pnpm_workspace_packages()
-    version_map = {pkg.name: pkg for pkg in packages}
-    for pkg in packages:
-        _sync_package_version_with_pyproject(pkg.path, version_map, pkg.name)
+    # Get the updated Python package version from py/package.json (updated by changesets)
+    py_package_path = Path("py/package.json")
+    if not py_package_path.exists():
+        click.echo("Python package.json not found", err=True)
+        sys.exit(1)
+
+    with open(py_package_path) as f:
+        py_package = json.load(f)
+
+    new_version = py_package["version"]
+    # Update Python pyproject.toml files based on the package.json version
+    update_python_versions(new_version)
+
+    click.echo(f"Successfully propagated version {new_version} to all Python packages")
 
 
 @cli.command()
 @click.option("--tag", is_flag=True, help="Tag the packages after publishing")
 @click.option("--dry-run", is_flag=True, help="Dry run the publish")
-@click.option("--js/--no-js", default=True, help="Publish the js package")
-@click.option("--py/--no-py", default=True, help="Publish the py package")
-def publish(tag: bool, dry_run: bool, js: bool, py: bool) -> None:
+def publish(tag: bool, dry_run: bool) -> None:
     """Publish all packages."""
     # move to the root
     os.chdir(Path(__file__).parent.parent)
 
-    if js:
-        if not os.getenv("NPM_TOKEN"):
-            click.echo("NPM_TOKEN is not set, skipping publish", err=True)
-            raise click.Abort("No token set")
-    if py:
-        if not os.getenv("LLAMA_PARSE_PYPI_TOKEN"):
-            click.echo("LLAMA_PARSE_PYPI_TOKEN is not set, skipping publish", err=True)
-            raise click.Abort("No token set")
+    if not os.getenv("NPM_TOKEN"):
+        click.echo("NPM_TOKEN is not set, skipping publish", err=True)
+        raise click.Abort("No token set")
+    if not os.getenv("LLAMA_PARSE_PYPI_TOKEN"):
+        click.echo("LLAMA_PARSE_PYPI_TOKEN is not set, skipping publish", err=True)
+        raise click.Abort("No token set")
 
     # not general script. Just checks each of the 2 packages to see if they need to be published.
-    if js:
-        maybe_publish_npm(dry_run)
-    if py:
-        maybe_publish_pypi(dry_run)
+    maybe_publish_ts_package(dry_run)
+    maybe_publish_py_packages(dry_run)
 
     if tag:
         if dry_run:
             click.echo("Dry run, skipping tag. Would run:")
             click.echo("  npx @changesets/cli tag")
             click.echo("  git push --tags")
+            return
         else:
-            # Let changesets create JS-related tags as usual
             _run_command(["npx", "@changesets/cli", "tag"])
             _run_command(["git", "push", "--tags"])
 
 
-def maybe_publish_npm(dry_run: bool) -> None:
+def maybe_publish_ts_package(dry_run: bool) -> None:
     """Publish the ts package if it needs to be published."""
     target_dir = Path("ts/llama_cloud_services")
     ts_path_package = target_dir / "package.json"
@@ -234,7 +169,7 @@ def maybe_publish_npm(dry_run: bool) -> None:
         _run_command(["pnpm", "publish"], cwd=target_dir)
 
 
-def maybe_publish_pypi(dry_run: bool) -> None:
+def maybe_publish_py_packages(dry_run: bool) -> None:
     """Publish the py packages if they need to be published."""
     for pyproject in list(Path("py").glob("*/pyproject.toml")) + [
         Path("py/pyproject.toml")