René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"[MNT]: Inconsistency in log-scale handling: `bbox_inches='tight'` creates massive canvas for negative text coordinates","articleBody":"### Summary\n\n### Problem Description\n\nWhen using `ax.set_yscale('log')`, Matplotlib correctly ignores/masks negative values in data series, and also Artists like text placed at a negative coordinate. However, when combined with `savefig(..., bbox_inches='tight')`, the layout engine attempts to include these \"invisible\" texts, resulting in a figure with a massive white canvas and a microscopic plot.\n\n\n\n### Minimal Reproducible Example\n\n### 1. The Normal Case (Data Masking)\n\nThis script shows that Matplotlib handles negative **data** gracefully. It simply doesn't draw the parts of the line that go below zero on a log scale.\n\n```python\nimport matplotlib.pyplot as plt\nimport numpy as np\n\n# Data goes from -10 to 10\nx = np.linspace(0, 10, 100)\ny = np.sin(x) * 10 \n\nfig, ax = plt.subplots()\nax.plot(x, y, label='Sine Wave')\nax.set_yscale('log')\nax.set_title(\"1. Data with negatives is masked correctly\")\n\nplt.savefig(\"1_normal_data.png\")\n\n\n```\n\n---\n\n### 2. The \"Hidden Danger\" (Negative Text)\n\nHere, we add text at negative places. Everything looks fine because the window is usually sized to the **axes**, not the **artists**. The text is simply \"off-screen\" in a mathematical void.\n\n```python\nimport matplotlib.pyplot as plt\nimport numpy as np\n\nx = np.linspace(0, 10, 100)\ny = np.sin(x) * 10 \n\nfig, ax = plt.subplots()\nax.plot(x, y)\nax.set_yscale('log')\n\n# Adding text at a negative Y-coordinate\nax.text(5, -5, \"I am at y = -5\", color='red', fontweight='bold')\nax.set_title(\"2. Text at y=-5 (Looks okay in viewer)\")\n\nplt.savefig(\"2_text_no_tight.png\")\n\n\n```\n\n---\n\n### 3. The \"Explosion\" (`bbox_inches='tight'`)\n\nThis is the script that triggers the issue. By using `bbox_inches='tight'`, you tell Matplotlib to \"shrink-wrap\" the image around **every** artist. Because it results in an extreme coordinate, the \"shrink-wrap\" stretches across thousands of virtual pixels.\n\n```python\nimport matplotlib.pyplot as plt\nimport numpy as np\n\nx = np.linspace(0, 10, 100)\ny = np.sin(x) * 10 \n\nfig, ax = plt.subplots()\nax.plot(x, y)\nax.set_yscale('log')\n\n# This text coordinate will 'hallucinate' a massive distance in log-space\nax.text(5, -5, \"I am at y = -5\", color='red')\nax.set_title(\"3. The Explosion (Check the saved file!)\")\n\n# This specific command triggers the canvas stretching\nplt.savefig(\"3_canvas_explosion.png\", bbox_inches='tight')\n\n\n```\n\n### Actual Behavior\n\nThe log transform of a negative number (e.g., ) results in a non-finite or extreme proxy value (maybe). The `tight_layout` / `bbox_inches` logic includes this coordinate in the bounding box calculation, stretching the figure canvas to extreme proportions.\n\n### Expected Behavior\n\nIdeally, Artists with mathematically undefined coordinates on the current scale should be treated similarly to data points:\n\n1. They should be excluded from the bounding box calculation for `bbox_inches='tight'`.\n2. Or, at least they should trigger a warning that an Artist is located at an undefined coordinate for the chosen scale, which may result in an unexpected figure.\n\n### Suggested Mitigation\n\nI know matplotlib provides tools like set_in_layout to control whether an artist is to be included in layout calculations, but `bbox_inches='tight'` originally tries to figure out the tight bbox of the figure [says the documentation], so I would suggest not considering those items when deciding bbox, or at least raise some warning in the early versions.\n\n\n\n### Proposed fix\n\nFor the modification, I would recommend adding some filters in [get_tightbbox of artist.py](https://github.com/matplotlib/matplotlib/blob/1392cbe3c79cdb93f9282747841d648770f60249/lib/matplotlib/artist.py#L348-L373) for situations of scale=='log'.\n\n[A part of expressions is refined by GPT but manually checked]","author":{"url":"https://github.com/leakyH","@type":"Person","name":"leakyH"},"datePublished":"2026-02-02T08:11:12.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":0},"url":"https://github.com/31054/matplotlib/issues/31054"}