Fix pydantic dependency conflict with llama-index-workflows

[Copilot]

Problem

PlanExe could not be installed with pip install .[gradio-ui] due to conflicting pydantic version requirements:

ERROR: Cannot install None, PlanExe, llama-index-core and planexe[gradio-ui]==2025.5.20 
because these package versions have conflicting dependencies.

The conflict is caused by:
    planexe 2025.5.20 depends on pydantic==2.10.4
    llama-index-workflows 1.3.0 depends on pydantic>=2.11.5

The root cause was that PlanExe pinned pydantic to exactly version 2.10.4, while llama-index-workflows (a transitive dependency of llama-index-core) requires pydantic >= 2.11.5. These constraints are mutually exclusive and prevented pip from resolving dependencies.

Solution

Updated pydantic version constraints in pyproject.toml to use version ranges that satisfy all dependencies:

• Changed: pydantic==2.10.4 → pydantic>=2.11.5,<3.0.0
• Changed: pydantic_core==2.27.2 → pydantic_core>=2.28.1

This new constraint satisfies all dependency requirements:

• ✅ llama-index-workflows: pydantic>=2.11.5
• ✅ llama-index-core: pydantic>=2.8.0
• ✅ fastapi 0.115.6: pydantic<3.0.0 and >=1.7.4
• ✅ gradio 5.16.0: pydantic>=2.0
• ✅ banks 2.2.0: pydantic (any version)

Changes

Core Fix

• pyproject.toml: Updated pydantic and pydantic_core constraints to use compatible version ranges

Verification & Documentation

• verify_installation.py: Added pydantic version check to validate installations
• verify_dependency_fix.py: New script to automatically verify dependency constraints are correct
• DEPENDENCY_FIX.md: Updated with comprehensive documentation of both pillow and pydantic fixes
• INSTALL.md: Added pydantic compatibility information

Testing

• ✅ All 20 existing unit tests pass
• ✅ CodeQL security scan: 0 alerts
• ✅ GitHub Advisory Database: No vulnerabilities in pydantic 2.11.5+
• ✅ Dependency verification script confirms all constraints are satisfied
• ✅ Mathematical verification of version compatibility

Why Version Ranges?

Using pydantic>=2.11.5,<3.0.0 instead of pinning to an exact version provides:

• Flexibility: Allows pip to choose any compatible version within the range
• Future compatibility: Automatically benefits from bug fixes and minor updates
• Better resolution: Makes it easier for pip to find solutions that satisfy all dependencies
• Standard practice: Follows Python packaging best practices

Impact

Users can now successfully install PlanExe with pip install .[gradio-ui] without dependency conflicts. The installation process is smooth and reliable, and the package remains compatible with the broader Python ecosystem.

Original prompt

ERROR: Cannot install None, PlanExe, llama-index-core and planexe[gradio-ui]==2025.5.20 because these package versions have conflicting dependencies.

The conflict is caused by:
planexe 2025.5.20 depends on pydantic==2.10.4
planexe[gradio-ui] 2025.5.20 depends on pydantic==2.10.4
fastapi 0.115.6 depends on pydantic!=1.8, !=1.8.1, !=2.0.0, !=2.0.1, !=2.1.0, <3.0.0 and >=1.7.4
gradio 5.16.0 depends on pydantic>=2.0
llama-index-core 0.13.0 depends on pydantic>=2.8.0
banks 2.2.0 depends on pydantic
llama-index-workflows 1.3.0 depends on pydantic>=2.11.5
llama-index-workflows 1.2.0 depends on pydantic>=2.11.5
llama-index-workflows 1.1.0 depends on pydantic>=2.11.5
llama-index-workflows 1.0.1 depends on pydantic>=2.11.5

To fix this you could try to:

1. loosen the range of package versions you've specified
2. remove package versions to allow pip to attempt to solve the dependency conflict

ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/topics/dependency-resolution/#dealing-with-dependency-conflicts

(venv) D:\Work\repo\PlanExe>

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.