Add concise macOS troubleshooting notes for common install/runtime issues.

.github/workflows/docker-publish.yml

@@ -73,7 +73,7 @@ jobs:
       # https://github.com/docker/build-push-action
       - name: Build and push Docker image
         id: build-and-push
-        uses: docker/build-push-action@0565240e2d4ab88bba5387d719585280857ece09 # v5.0.0
+        uses: docker/build-push-action@263435318d21b8e681c14492fe198d362a7d2c83 # v6.18.0
         with:
           context: .
           push: ${{ github.event_name != 'pull_request' }}

.gitignore

@@ -101,3 +101,5 @@ __pypackages__/
 
 # Added by cargo
 /target
+
+docs/installation-troubleshooting/_generated_index.md

docs/installation-troubleshooting/index.md

@@ -0,0 +1,92 @@
+# Installation Troubleshooting
+
+This directory contains concise, platform-specific troubleshooting notes focused on reproducible diagnostics and resolutions.
+
+Files:
+
+- `macos.md` — macOS troubleshooting index and links to per-entry files.
+- `macos-poppler-build.md` — Poppler build failure entry.
+- `macos-poppler-runtime.md` — Poppler runtime binaries entry.
+- `macos-tesseract.md` — Tesseract / pytesseract entry.
+- `linux.md` — placeholder for Linux troubleshooting entries.
+- `windows.md` — placeholder for Windows troubleshooting entries.
+
+Contribution guidance
+
+When adding a troubleshooting entry, include the following fields to make the entry reproducible and searchable:
+
+- Title: short, descriptive title.
+- Symptom: exact error message or observable behavior.
+- Environment: OS, architecture (x86_64 / arm64), Python version, Pixi version (if relevant), and any other relevant software versions.
+- Reproduction steps: exact commands or minimal steps to reproduce the issue.
+- Resolution: exact commands and a brief explanation of why the resolution works.
+- References: links to upstream docs, issues, or relevant resources.
+
+
+Format example (minimal template with YAML frontmatter for new troubleshooting entries):
+
+```yaml
+---
+title: "Brief title"
+date: "YYYY-MM-DD"       # Use YYYY-MM-DD (e.g., "2025-08-28")
+verified_on: null        # Set to validation date (e.g., "2025-08-29") after verification
+os: macos                # single value or a list: ["macos", "linux", "windows"]
+arch: any                # one of: "x86_64", "arm64", "any"
+severity: low            # one of: "low", "medium", "high"
+status: active           # "active" or "deprecated"
+tags: [technology, error-type]  # e.g., [poppler, pdf2image]
+reproducible: true
+---
+```
+
+Symptom: copy-paste of error/output
+Environment: e.g. macOS 13.5, arm64, Python 3.12, pixi 0.1
+Reproduction steps:
+  1. command1
+  2. command2
+Resolution:
+  - command(s)
+  - short explanation
+References: link(s)
+
+Notes:
+
+- Keep entries concise and diagnostic-first; avoid duplicating general installation instructions already present in `README.rst` or Pixi docs.
+- Per-OS files for clarity and searchability. If an entry is cross-platform, note that in the Environment field and add it to each relevant file.
+
+Maintenance guidance:
+
+- Each entry should include `date` and, when possible, `verified_on` to indicate when it was validated. Maintainers should mark `status: deprecated` if an entry becomes obsolete and point to the canonical install docs.
+- Automated checks should verify frontmatter presence and basic fields.
+
+## Docs tooling
+
+The repository provides a small local tool to validate troubleshooting entries and generate a consolidated index. This tool is optional and non-blocking by default.
+
+Run the validator and index generator inside the project's Pixi environment:
+
+```bash
+pixi run python3 tools/generate_troubleshooting_index.py
+```
+
+Use `--strict` to make the tool exit non-zero when required frontmatter is missing (useful for CI once the team decides to enforce metadata):
+
+```bash
+pixi run python3 tools/generate_troubleshooting_index.py --strict
+```
+
+Minimal frontmatter template for an entry (keep entries short and link to canonical README installation instructions):
+
+```yaml
+---
+title: "Short descriptive title"
+date: 2025-08-28
+os: macos
+arch: any
+status: active
+tags: [poppler, pdf2image]
+---
+```
+
+The tool writes `docs/installation-troubleshooting/_generated_index.md`. You do not need to commit the generated file; it can be regenerated on demand. For now, the script will only warn about missing metadata unless `--strict` is used.
+

docs/installation-troubleshooting/linux.md

@@ -0,0 +1,10 @@
+# Linux troubleshooting (landing)
+
+This page points to Linux-specific troubleshooting files. Create per-issue markdown files in this directory and follow the frontmatter template in `index.md`.
+
+Examples to consider:
+
+- package manager issues (apt/yum) for Poppler/Tesseract
+- Python wheel compatibility (GLIBC/C++ ABI) problems
+
+Add new entries as focused, reproducible files; avoid duplicating canonical installation docs.

docs/installation-troubleshooting/macos-pixi-conda-pypi-warning.md

@@ -0,0 +1,45 @@
+---
+title: "Pixi warning: conda-packages overridden by PyPI"
+date: "2025-08-28"
+verified_on: "2025-08-28"
+os: macos
+arch: any
+severity: low
+status: active
+tags: [pixi, conda, pip, environment]
+reproducible: true
+---
+
+Symptom
+-------
+
+When running `pixi shell -e pdev`, you may see a warning such as:
+
+```
+WARN These conda-packages will be overridden by pypi:
+   yarg
+```
+
+Environment
+-----------
+
+- Using Pixi to create or enter a project environment that mixes Conda and PyPI package sources.
+
+Reproduction steps
+------------------
+
+1. Run `pixi shell -e pdev` (or the equivalent Pixi environment creation command).
+2. Observe the warning during environment resolution.
+
+Resolution
+----------
+
+This warning indicates that a package available from both Conda and PyPI will be installed from PyPI instead of the Conda channel. Usually this is safe, but if you rely on the Conda-built binary (for example, for performance or compatibility), align the package source by:
+
+- Pinning the package to the Conda channel in the Pixi/Pixienv configuration.
+- Removing the PyPI override from the project's requirements if Conda should be authoritative.
+
+Notes
+-----
+
+Most users can safely ignore this warning unless they encounter runtime issues tied to a specific build of the package.

docs/installation-troubleshooting/macos-poppler-build.md

@@ -0,0 +1,77 @@
+---
+title: "Poppler build failure: missing macOS SDK headers"
+date: "2025-08-28"
+verified_on: "2025-08-28"
+os: macos
+arch: arm64
+severity: medium
+status: active
+tags: [poppler, build, xcode]
+reproducible: true
+---
+
+Symptom
+-------
+
+Compilation fails with errors like:
+
+```
+fatal error: 'ctime' file not found
+```
+
+or other C/C++ header errors when installing packages that build native extensions (for example, Poppler utilities or bindings).
+
+Environment
+-----------
+
+- macOS host (any recent release)
+- Building Python packages that include C/C++ extensions
+
+Reproduction steps
+------------------
+
+1. Create or update the Pixi environment that installs packages requiring native builds (or run `pip install` for such packages).
+2. Observe the build failure during the compilation step.
+
+Resolution
+----------
+
+1. Install or update the Xcode Command Line Tools:
+
+```bash
+xcode-select --install
+```
+
+2. If the system points to the wrong developer directory, reset it:
+
+```bash
+sudo xcode-select --reset
+```
+
+3. If prompted, accept the Xcode license:
+
+```bash
+sudo xcodebuild -license
+```
+
+4. If the build still cannot find headers, export a valid SDK path for the build process and retry:
+
+```bash
+export SDKROOT=$(xcrun --sdk macosx --show-sdk-path)
+# then re-run the install or build
+```
+
+Notes
+-----
+
+These steps provide the compiler and headers required to build C/C++ extensions. Prefer installing Poppler via the OS package manager (Homebrew) when possible to avoid local compilation.
+
+References
+----------
+
+- https://developer.apple.com
+
+Verified on
+----------
+
+- Apple M1 Pro (2021), macOS 15.6.1 (verification date: 2025-08-28)

docs/installation-troubleshooting/macos-poppler-runtime.md

@@ -0,0 +1,62 @@
+---
+title: "Poppler runtime missing binaries"
+date: "2025-08-28"
+verified_on: "2025-08-28"
+os: macos
+arch: arm64
+severity: medium
+status: active
+tags: [poppler, pdf2image, runtime]
+reproducible: true
+---
+
+Symptom
+-------
+
+Runtime errors indicate Poppler utilities are missing, for example:
+
+```
+FileNotFoundError: [Errno 2] No such file or directory: 'pdftoppm'
+```
+
+Environment
+-----------
+
+- macOS host where Python calls `pdf2image` or other Poppler-dependent tooling
+
+Reproduction steps
+------------------
+
+1. Run a PDF conversion that relies on Poppler (e.g., code that uses `pdf2image`).
+2. Observe the FileNotFoundError or similar runtime error.
+
+Resolution
+----------
+
+Install Poppler via Homebrew:
+
+```bash
+brew install poppler
+```
+
+Verify the binaries are on PATH:
+
+```bash
+which pdftoppm
+which pdftotext
+```
+
+Notes
+-----
+
+`pdf2image` requires host-installed Poppler binaries. Pixi manages Python packages but not host-level system packages; install the latter with Homebrew, apt, or your OS package manager.
+
+References
+----------
+
+- https://poppler.freedesktop.org/
+
+Verified on
+----------
+
+- Apple M1 Pro (2021), macOS 15.6.1 (verification date: 2025-08-28)

docs/installation-troubleshooting/macos-syntaxwarning.md

@@ -0,0 +1,59 @@
+---
+title: "SyntaxWarning: invalid escape sequence in crawl4ai/prompts.py"
+date: "2025-08-28"
+verified_on: "2025-08-28"
+os: macos
+arch: arm64
+severity: low
+status: active
+tags: [python, warning, syntax]
+reproducible: true
+---
+
+Symptom
+-------
+
+You may see a warning like:
+
+```
+SyntaxWarning: invalid escape sequence '\`'
+```
+
+Environment
+-----------
+
+- A Python file in the repository (example: `crawl4ai/prompts.py`) contains a string with a backslash followed by a character that doesn't form a valid escape sequence.
+
+Reproduction steps
+------------------
+
+1. Run code that imports or executes the affected module (or run linting/tests).
+2. Observe the SyntaxWarning printed at import time or during execution.
+
+Resolution
+----------
+
+Update the affected string to use a valid escape or a raw string. For example, either double-escape the backslash:
+
+```py
+# before
+text = "This contains a backslash and a backtick: \`"
+
+# after
+text = "This contains a backslash and a backtick: \\\`"
+```
+
+or mark the string as raw (if that semantics is acceptable):
+
+```py
+text = r"This contains a backslash and a backtick: \`"
+```
+
+Notes
+-----
+
+This is a low-severity, linter-level issue; it won't usually break runtime behavior but fixing it keeps console output clean and avoids noisy CI logs.
+
+References
+----------
+- Python string literal docs: https://docs.python.org/3/reference/lexical_analysis.html#string-and-bytes-literals