chore: version packages (#965)

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

.github/workflows/build_package_ts.yml

@@ -21,7 +21,7 @@ jobs:
       - uses: pnpm/action-setup@v4
 
       - name: Setup Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v5
         with:
           node-version-file: "ts/llama_cloud_services/.nvmrc"
 

.github/workflows/lint.yml

@@ -31,7 +31,7 @@ jobs:
 
       - uses: pnpm/action-setup@v4
       - name: Setup Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v5
         with:
           node-version-file: "ts/llama_cloud_services/.nvmrc"
       - name: Install dependencies

.github/workflows/test_ts.yml

@@ -24,7 +24,7 @@ jobs:
       - uses: actions/checkout@v5
       - uses: pnpm/action-setup@v4
       - name: Setup Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v5
         with:
           node-version-file: "ts/llama_cloud_services/.nvmrc"
       - name: Install dependencies

.github/workflows/version_bump_and_release.yml

@@ -15,23 +15,23 @@ jobs:
     if: github.ref == 'refs/heads/main'
     steps:
       - name: Checkout Repo
-        uses: actions/checkout@v4
+        uses: actions/checkout@v5
 
-      - uses: pnpm/action-setup@v3
+      - uses: pnpm/action-setup@v4
 
       - name: Setup Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v5
         with:
           node-version: "22"
           cache: "pnpm"
 
       - name: Setup Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
           python-version: "3.11"
 
       - name: Install uv
-        uses: astral-sh/setup-uv@v3
+        uses: astral-sh/setup-uv@v6
 
       - name: Install dependencies
         run: pnpm install

package.json

@@ -8,7 +8,7 @@
   "scripts": {
     "pre-commit-version": "pnpm changeset",
     "version": "./scripts/changeset-version.py version",
-    "publish": "./scripts/changeset-version.py publish"
+    "publish": "./scripts/changeset-version.py publish --tag"
   },
   "devDependencies": {
     "prettier": "^3.6.2",

pnpm-lock.yaml

@@ -21,7 +21,21 @@ importers:
         specifier: ^3.6.2
         version: 3.6.2
 
-  py: {}
+  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
 
   ts/e2e-tests:
     devDependencies:

pnpm-workspace.yaml

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

py/CHANGELOG.md

@@ -1,5 +1,30 @@
 # llama-cloud-services-py
 
+## 0.6.74
+
+### Patch Changes
+
+- 1bf5223: Fix default bbox values
+- 24166dc: Now only escape single dollar signs - preserve double for latex equations
+
+## 0.6.73
+
+### Patch Changes
+
+- e6a7939: Loosen packaging dep requirement
+
+## 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

@@ -4,7 +4,10 @@ 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
+from llama_cloud_services.parse.utils import (
+    make_api_request,
+    is_jupyter,
+)
 from llama_index.core.async_utils import asyncio_run
 from llama_index.core.schema import Document, ImageDocument, ImageNode, TextNode
 
@@ -27,10 +30,22 @@ class JobMetadata(BaseModel):
 class BBox(BaseModel):
     """A bounding box."""
 
-    x: float = Field(description="The x-coordinate of the bounding box.")
-    y: float = Field(description="The y-coordinate of the bounding box.")
-    w: float = Field(description="The width of the bounding box.")
-    h: float = Field(description="The height of the bounding box.")
+    x: Optional[float] = Field(
+        default=None,
+        description="The x-coordinate of the bounding box.",
+    )
+    y: Optional[float] = Field(
+        default=None,
+        description="The y-coordinate of the bounding box.",
+    )
+    w: Optional[float] = Field(
+        default=None,
+        description="The width of the bounding box.",
+    )
+    h: Optional[float] = Field(
+        default=None,
+        description="The height of the bounding box.",
+    )
 
 
 class PageItem(BaseModel):
@@ -85,12 +100,16 @@ class LayoutItem(BaseModel):
     """The layout of a page."""
 
     image: str = Field(description="The name of the image containing the layout item")
-    confidence: float = Field(description="The confidence of the layout item.")
+    confidence: float = Field(
+        default=0.0, description="The confidence of the layout item."
+    )
     label: str = Field(description="The label of the layout item.")
     bbox: Optional[BBox] = Field(
         default=None, description="The bounding box of the layout item."
     )
-    isLikelyNoise: bool = Field(description="Whether the layout item is likely noise.")
+    isLikelyNoise: bool = Field(
+        default=False, description="Whether the layout item is likely noise."
+    )
 
 
 class ChartItem(BaseModel):
@@ -258,6 +277,29 @@ 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_single_dollar_signs(text: str) -> str:
+            """Escape single dollar signs in text to prevent Jupyter from interpreting them as LaTeX.
+
+            Preserves all strings of dollar signs greater than length 1,
+            especially preserving double dollar signs ($$) which denote LaTeX equations.
+
+            Args:
+                text: The text to escape
+
+            Returns:
+                Text with single dollar signs escaped
+            """
+            # Replace single $ with \$, but preserve $$
+            # Use negative lookahead and lookbehind to match $ not preceded or followed by $
+            return re.sub(r"(?<!\$)\$(?!\$)", r"\$", text)
+
+        return escape_single_dollar_signs(text)
+
     def get_markdown_documents(self, split_by_page: bool = False) -> List[Document]:
         """
         Get the markdown documents from the job.
@@ -268,17 +310,22 @@ class JobResult(BaseModel):
         if split_by_page:
             return [
                 Document(
-                    text=page.md,
+                    text=self._format_markdown_for_notebook(page.md)
+                    if is_jupyter()
+                    else 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._page_separator.join(
-                        [page.md if page.md is not None else "" for page in self.pages]
-                    ),
+                    text=self._format_markdown_for_notebook(text)
+                    if is_jupyter()
+                    else text,
                     metadata={"file_name": self.file_name},
                 )
             ]
@@ -328,7 +375,10 @@ 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)
-        return response.content.decode("utf-8")
+        markdown = response.content.decode("utf-8")
+        return (
+            self._format_markdown_for_notebook(markdown) if is_jupyter() else markdown
+        )
 
     def get_text(self) -> str:
         """

py/llama_cloud_services/parse/utils.py

@@ -1,3 +1,4 @@
+import functools
 import httpx
 import itertools
 import logging
@@ -356,6 +357,17 @@ 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

@@ -0,0 +1,23 @@
+# llama_parse
+
+## 0.6.74
+
+### Patch Changes
+
+- Updated dependencies [1bf5223]
+- Updated dependencies [24166dc]
+  - llama-cloud-services-py@0.6.74
+
+## 0.6.73
+
+### Patch Changes
+
+- Updated dependencies [e6a7939]
+  - llama-cloud-services-py@0.6.73
+
+## 0.6.72
+
+### Patch Changes
+
+- Updated dependencies [ad6734b]
+  - llama-cloud-services-py@0.6.72

py/llama_parse/package.json

@@ -0,0 +1,20 @@
+{
+  "name": "llama_parse",
+  "version": "0.6.74",
+  "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 = "0.6.70"
+version = "0.6.74"
 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>=0.6.70"]
+dependencies = ["llama-cloud-services>=0.6.74"]
 
 [project.scripts]
 llama-parse = "llama_parse.cli.main:parse"

py/package.json

@@ -1,7 +1,10 @@
 {
   "name": "llama-cloud-services-py",
-  "version": "0.6.70",
-  "private": "true",
+  "version": "0.6.74",
+  "private": false,
   "license": "MIT",
-  "scripts": {}
+  "scripts": {},
+  "devDependencies": {
+    "changesets": "^1.0.2"
+  }
 }

py/pyproject.toml

@@ -19,7 +19,7 @@ dev = [
 
 [project]
 name = "llama-cloud-services"
-version = "0.6.70"
+version = "0.6.74"
 description = "Tailored SDK clients for LlamaCloud services."
 authors = [{name = "Logan Markewich", email = "logan@runllama.ai"}]
 requires-python = ">=3.9,<4.0"
@@ -34,7 +34,7 @@ dependencies = [
   "eval-type-backport>=0.2.0,<0.3 ; python_version < '3.10'",
   "platformdirs>=4.3.7,<5",
   "tenacity>=8.5.0, <10.0",
-  "packaging>=25.0"
+  "packaging>=23.0"
 ]
 
 [project.scripts]

py/tests/parse/test_llama_parse_result.py

@@ -6,6 +6,40 @@ from llama_cloud_services import LlamaParse
 from llama_cloud_services.parse.types import JobResult
 
 
+def test_format_parse_result_markdown_for_notebook():
+    """Test the _format_markdown_for_notebook function.
+    Right now, the only work it does is escape single dollar signs."""
+    result = JobResult(job_id="test", file_name="test.pdf", job_result={})
+
+    # Test None input
+    assert result._format_markdown_for_notebook(None) is None
+
+    # Test single dollar sign gets escaped
+    assert result._format_markdown_for_notebook("This costs $5") == "This costs \\$5"
+
+    # Test double dollar signs are preserved (LaTeX equations)
+    assert (
+        result._format_markdown_for_notebook("$$x^2 + y^2 = z^2$$")
+        == "$$x^2 + y^2 = z^2$$"
+    )
+
+    # Test mixed single and double dollar signs
+    text = "This costs $5, but $$E = mc^2$$ is priceless"
+    expected = "This costs \\$5, but $$E = mc^2$$ is priceless"
+    assert result._format_markdown_for_notebook(text) == expected
+
+    # Test multiple single dollar signs
+    assert result._format_markdown_for_notebook("$10 and $20") == "\\$10 and \\$20"
+
+    # Test three or more consecutive dollar signs (preserve them)
+    assert result._format_markdown_for_notebook("$$$") == "$$$"
+
+    # Test adjacent dollar signs with text in between
+    text = "$$inline$$ and $separate"
+    expected = "$$inline$$ and \\$separate"
+    assert result._format_markdown_for_notebook(text) == expected
+
+
 @pytest.fixture
 def file_path() -> str:
     return "tests/test_files/attention_is_all_you_need.pdf"

py/uv.lock

@@ -1,5 +1,5 @@
 version = 1
-revision = 2
+revision = 3
 requires-python = ">=3.9, <4.0"
 resolution-markers = [
     "python_full_version >= '3.14'",
@@ -1596,7 +1596,7 @@ wheels = [
 
 [[package]]
 name = "llama-cloud-services"
-version = "0.6.69"
+version = "0.6.73"
 source = { editable = "." }
 dependencies = [
     { name = "click", version = "8.1.8", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.10'" },
@@ -1633,7 +1633,7 @@ requires-dist = [
     { name = "eval-type-backport", marker = "python_full_version < '3.10'", specifier = ">=0.2.0,<0.3" },
     { name = "llama-cloud", specifier = "==0.1.43" },
     { name = "llama-index-core", specifier = ">=0.12.0" },
-    { name = "packaging", specifier = ">=25.0" },
+    { name = "packaging", specifier = ">=23.0" },
     { name = "platformdirs", specifier = ">=4.3.7,<5" },
     { name = "pydantic", specifier = ">=2.8,!=2.10" },
     { name = "python-dotenv", specifier = ">=1.0.1,<2" },

scripts/changeset-version.py

@@ -12,14 +12,15 @@ 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 List
+from typing import Any, List, cast
 import urllib.request
 import urllib.error
+import re
 
 import click
 import tomlkit
@@ -33,40 +34,97 @@ def _run_command(
     subprocess.run(cmd, check=True, text=True, cwd=cwd or Path.cwd(), env=env)
 
 
-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.
+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
 
-    This function updates the version in both pyproject.toml files.
+
+@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.
     """
-    # 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))
+    pyproject_path = package_dir / "pyproject.toml"
+    if not pyproject_path.exists():
+        return
 
-    # 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))
+    package_version = packages[js_package_name].version
+    py_doc = tomlkit.parse(pyproject_path.read_text())
 
-    # 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
+    by_python_name = {
+        pkg.python_package_name(): pkg
+        for pkg in packages.values()
+        if pkg.python_package_name()
+    }
 
-    parse_path.write_text(tomlkit.dumps(parse_doc))
+    current_version = py_doc["project"]["version"]
+    assert isinstance(current_version, str)
 
-    click.echo(f"Updated Python packages to version {version}")
+    # 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"
+        )
 
 
 def lock_python_dependencies() -> None:
@@ -86,57 +144,64 @@ def cli() -> None:
 
 @cli.command()
 def version() -> None:
-    """Apply changeset versions, and propagate them to Python packages."""
-    # First, run changeset version to update all package.json files (including py/package.json)
+    """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
     _run_command(["npx", "@changesets/cli", "version"])
 
-    # 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")
+    # 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)
 
 
 @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")
-def publish(tag: bool, dry_run: bool) -> None:
+@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:
     """Publish all packages."""
     # move to the root
     os.chdir(Path(__file__).parent.parent)
 
-    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")
+    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")
 
     # not general script. Just checks each of the 2 packages to see if they need to be published.
-    maybe_publish_ts_package(dry_run)
-    maybe_publish_py_packages(dry_run)
+    if js:
+        maybe_publish_npm(dry_run)
+    if py:
+        maybe_publish_pypi(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_ts_package(dry_run: bool) -> None:
+def maybe_publish_npm(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"
@@ -169,7 +234,7 @@ def maybe_publish_ts_package(dry_run: bool) -> None:
         _run_command(["pnpm", "publish"], cwd=target_dir)
 
 
-def maybe_publish_py_packages(dry_run: bool) -> None:
+def maybe_publish_pypi(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")

ts/llama_cloud_services/CHANGELOG.md

@@ -1,5 +1,11 @@
 # llama-cloud-services
 
+## 0.3.8
+
+### Patch Changes
+
+- 6e0f2f4: Agent data extraction citations can be undefined
+
 ## 0.3.7
 
 ### Patch Changes

ts/llama_cloud_services/package.json

@@ -1,6 +1,6 @@
 {
   "name": "llama-cloud-services",
-  "version": "0.3.7",
+  "version": "0.3.8",
   "type": "module",
   "license": "MIT",
   "scripts": {

ts/llama_cloud_services/src/beta/agent/types.ts

@@ -38,7 +38,7 @@ export interface ExtractedFieldMetadata {
   confidence?: number;
   /** The confidence score for the field based on the extracted text only */
   extraction_confidence?: number;
-  citation: FieldCitation[];
+  citation?: FieldCitation[];
 }
 
 export interface FieldCitation {