⚡️ Speed up function standalone_docs_json_and_render_items by 8%

[codeflash-ai]

📄 8% (0.08x) speedup for standalone_docs_json_and_render_items in src/bokeh/embed/util.py

⏱️ Runtime : 6.38 milliseconds → 5.90 milliseconds (best of 59 runs)

📝 Explanation and details

The optimized code achieves an 8% speedup through several targeted micro-optimizations that reduce function call overhead and improve algorithmic efficiency:

Key Optimizations:

1. Dict access optimization in standalone_docs_json_and_render_items: Replaced if doc not in docs: check followed by assignment with docs.setdefault(). This eliminates redundant hash table lookups, saving ~0.5% of total runtime.

2. Short-circuit evaluation in submodel_has_python_callbacks: Replaced explicit loop with early-break logic using any() generator expression. This stops iteration immediately when the first model with callbacks is found, avoiding unnecessary model traversal.

3. List comprehension for render items: Changed from appending items in a loop to a single list comprehension, reducing function call overhead and temporary variable assignments.

4. Eliminated redundant imports in serialization functions: Moved the from ..core.types import ID import out of make_globally_unique_id(), removing repeated import overhead that was consuming ~18% of the function's time according to the profiler.

5. Optimized UUID generation in make_globally_unique_css_safe_id: Direct string generation from uuid.uuid4() instead of calling make_globally_unique_id() and converting back, reducing function call overhead.

6. Document model filtering improvement: Changed list comprehension to explicit loop for data_models collection, which shows better performance characteristics in the profiler (63.5% vs 65.9% of to_json time).

Impact on Hot Paths:
The function is called from notebook_content() in Jupyter notebook rendering, making these optimizations valuable for interactive data science workflows where plots are frequently generated. The 8% improvement translates to faster notebook cell execution and better user experience when creating multiple visualizations.

Test Case Performance:
The optimizations show consistent 5-13% improvements across error handling test cases, indicating the changes maintain correctness while improving performance across different code paths.

✅ Correctness verification report:

Test	Status
⚙️ Existing Unit Tests	✅ 44 Passed
🌀 Generated Regression Tests	✅ 7 Passed
⏪ Replay Tests	🔘 None Found
🔎 Concolic Coverage Tests	🔘 None Found
📊 Tests Coverage	100.0%

⚙️ Existing Unit Tests and Runtime

Test File::Test Function	Original ⏱️	Optimized ⏱️	Speedup
unit/bokeh/embed/test_util__embed.py::Test_standalone_docs_json.test_output	1.12ms	947μs	18.2%✅
unit/bokeh/embed/test_util__embed.py::Test_standalone_docs_json_and_render_items.test_exception_for_missing_doc	24.8μs	24.7μs	0.287%✅
unit/bokeh/embed/test_util__embed.py::Test_standalone_docs_json_and_render_items.test_log_warning_if_python_event_callback	2.13ms	2.02ms	5.29%✅
unit/bokeh/embed/test_util__embed.py::Test_standalone_docs_json_and_render_items.test_log_warning_if_python_property_callback	2.12ms	2.02ms	4.67%✅
unit/bokeh/embed/test_util__embed.py::Test_standalone_docs_json_and_render_items.test_suppress_warnings	953μs	850μs	12.2%✅

🌀 Generated Regression Tests and Runtime

import logging
# Helper for unique IDs
import uuid
from typing import Sequence

# imports
import pytest
from bokeh.embed.util import standalone_docs_json_and_render_items

# --- Minimal stubs/mocks for Bokeh core classes and helpers ---

class RenderItem:
    # Minimal stub for RenderItem used in the function
    def __init__(self, docid, roots):
        self.docid = docid
        self.roots = roots

    def __eq__(self, other):
        return isinstance(other, RenderItem) and self.docid == other.docid and self.roots == other.roots

    def __repr__(self):
        return f"RenderItem(docid={self.docid!r}, roots={self.roots!r})"

class Model:
    # Minimal stub for Bokeh Model
    def __init__(self, name=None):
        self.name = name
        self.document = None
        self._callbacks = []
        self._event_callbacks = []

    def __repr__(self):
        return f"Model({self.name!r})"

_CALLBACKS_WARNING = "You are generating standalone HTML/JS output, but trying to use real Python callbacks"
from bokeh.embed.util import standalone_docs_json_and_render_items

# --- Unit tests ---

# 1. Basic Test Cases






def test_model_not_in_document_raises():
    # Test that passing a model not attached to a document raises ValueError
    m = Model("orphan")
    with pytest.raises(ValueError):
        standalone_docs_json_and_render_items(m) # 4.36μs -> 4.15μs (5.16% faster)


def test_invalid_input_type_raises():
    # Test that passing an invalid input type raises ValueError
    with pytest.raises(ValueError):
        standalone_docs_json_and_render_items(123) # 4.12μs -> 4.05μs (1.70% faster)
    with pytest.raises(ValueError):
        standalone_docs_json_and_render_items([Model("a"), 42]) # 3.26μs -> 2.98μs (9.21% faster)

import logging
import uuid
from typing import Any, Sequence

# imports
import pytest
from bokeh.embed.util import standalone_docs_json_and_render_items

# --- Minimal stubs for Bokeh types and utilities ---

# Simulate Bokeh's ID type
class ID(str):
    pass

# Simulate Bokeh's DocJson type
class DocJson(dict):
    pass

# Simulate Bokeh's Model base class
class Model:
    def __init__(self, name=None):
        self.name = name
        self.document = None
        self._callbacks = []
        self._event_callbacks = []

# Simulate Bokeh's RenderItem class
class RenderItem:
    def __init__(self, docid, roots):
        self.docid = docid
        self.roots = roots

    def __eq__(self, other):
        return (
            isinstance(other, RenderItem)
            and self.docid == other.docid
            and self.roots == other.roots
        )

# --- Function under test ---

_CALLBACKS_WARNING = "PYTHON CALLBACKS WARNING"
from bokeh.embed.util import standalone_docs_json_and_render_items

# --- Unit tests ---

# Basic Test Cases





def test_model_without_document_edge():
    # Model not attached to any document should raise ValueError
    m = Model(name="orphan")
    with pytest.raises(ValueError):
        standalone_docs_json_and_render_items(m) # 4.33μs -> 4.08μs (5.97% faster)

def test_invalid_input_type_edge():
    # Passing an invalid type should raise ValueError
    with pytest.raises(ValueError):
        standalone_docs_json_and_render_items(123) # 3.48μs -> 3.10μs (12.3% faster)
    with pytest.raises(ValueError):
        standalone_docs_json_and_render_items([Model(name="x"), 123]) # 3.12μs -> 2.76μs (13.1% faster)

def test_empty_sequence_edge():
    # Empty sequence should not raise, but produce empty output
    docs_json, render_items = standalone_docs_json_and_render_items([]) # 11.4μs -> 11.5μs (0.767% slower)

To edit these changes git checkout codeflash/optimize-standalone_docs_json_and_render_items-mhwtcl1l and push.