Skip to main content

AI-powered CLI tool that migrates Python code to handle breaking dependency changes

Project description

PyResolve

CI PyPI version Python 3.9+ License: MIT

Don't just flag the update. Fix the break.

PyResolve is an AI-powered CLI tool that migrates Python code to handle breaking dependency changes. Unlike Dependabot/Renovate which just bump versions, PyResolve actually rewrites code to be compatible with new library versions.

Features

  • Auto-generated knowledge bases - Fetches changelogs and migration guides from GitHub, parses them with LLM to detect breaking changes
  • Tiered migration approach - Deterministic AST transforms for known patterns, KB-guided LLM for medium confidence, pure LLM fallback for complex cases
  • Confidence-based change detection - Shows HIGH/MEDIUM/LOW confidence breaking changes before migration
  • Local test execution to validate changes before applying
  • Beautiful diff output with explanations for each change

Supported Libraries

Tier 1 Libraries (Deterministic AST Transforms)

Library Migration Path Status
Pydantic v1 → v2 ✅ Full support
FastAPI 0.x → 0.100+ ✅ Supported
SQLAlchemy 1.4 → 2.0 ✅ Supported
Pandas 1.x → 2.x ✅ Supported
Requests Various ✅ Supported

Any Library (Auto-Generated Knowledge Base)

PyResolve can migrate any Python library by automatically fetching changelogs from GitHub and detecting breaking changes. For libraries not in Tier 1, it uses KB-guided or pure LLM migration.

Installation

pip install pyresolve

Quick Start

# Scan your project for all possible migrations
pyresolve scan

# Scan with detailed breaking change analysis
pyresolve scan --fetch-changes

# Analyze and propose changes for a specific library
pyresolve upgrade pydantic --target 2.5.0

# View detailed diff of proposed changes
pyresolve diff

# Apply changes to your files
pyresolve apply

Example Output

$ pyresolve upgrade pydantic --target 2.5.0

╭──────────────────────── PyResolve Migration ─────────────────────────╮
│ Upgrading Pydantic to version 2.5.0                                  │
│ Migration guide: https://docs.pydantic.dev/latest/migration/         │
╰──────────────────────────────────────────────────────────────────────╯

Fetching knowledge sources...
    GitHub: CHANGELOG.md
    GitHub: docs/migration.md

Breaking changes detected:

   HIGH CONFIDENCE:
   ├── .dict()  .model_dump()
   ├── @validator  @field_validator
   └── class Config  model_config = ConfigDict()

   MEDIUM CONFIDENCE:
   ├── .json()  .model_dump_json()
   └── parse_obj()  model_validate()

Scanning for library usage...
Found 12 imports from pydantic
Found 45 usages of pydantic symbols

┏━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━┓
┃ File                Changes  Status ┃
┡━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━┩
│ src/models/user.py        5  Ready  │
│ src/api/schemas.py        3  Ready  │
└────────────────────┴─────────┴────────┘

Total: 8 changes across 2 files

Usage

Scan Command

Scan your entire project for possible dependency migrations:

pyresolve scan

# Options:
#   --path, -p         Path to scan (default: current directory)
#   --fetch-changes    Fetch changelogs and detect breaking changes (slower)
#   --major-only       Only show major version upgrades
#   --json-output      Output results as JSON
#   --verbose, -v      Show detailed output

Example output:

$ pyresolve scan

Found 13 dependencies

Outdated Dependencies (5)
┏━━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━┳━━━━━━━┳━━━━━━━━━━┓
┃ Package     Current  Latest  Type     Tier   ┃
┡━━━━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━╇━━━━━━━╇━━━━━━━━━━┩
│ pydantic        1.0   2.5.0  Major   Tier 1  │
│ rich           13.0  14.0.0  Major  Tier 2/3 │
└────────────┴─────────┴────────┴───────┴──────────┘

Suggested Migrations (2)
  pydantic 1.0  2.5.0 (Tier 1 - deterministic)
  rich 13.0  14.0.0 (Tier 2/3 - LLM-assisted)

Quick commands:
  pyresolve upgrade pydantic --target 2.5.0
  pyresolve upgrade rich --target 14.0.0

Upgrade Command

Analyze your codebase and propose changes for a library upgrade:

pyresolve upgrade <library> --target <version>

# Options:
#   --target, -t    Target version to upgrade to
#   --path, -p      Path to analyze (default: current directory)
#   --dry-run       Show what would be changed without saving state

Diff Command

View the detailed diff of proposed changes:

pyresolve diff

# Options:
#   --file, -f      Show diff for specific file only
#   --no-color      Disable colored output

Apply Command

Apply the proposed changes to your files:

pyresolve apply

# Options:
#   --backup        Create backup files before applying changes
#   --file, -f      Apply changes to specific file only

How It Works

┌─────────────────────────────────────────────────────────────────────┐
│                     Knowledge Acquisition Pipeline                   │
│  ┌─────────────┐   ┌──────────────────┐   ┌─────────────────────┐  │
│  │ Local Cache │──▶│ On-Demand Gen    │──▶│ LLM Parser          │  │
│  │ (instant)   │   │ (fetches sources)│   │ (breaking changes)  │  │
│  └─────────────┘   └──────────────────┘   └─────────────────────┘  │
│                            │                                        │
│            ┌───────────────┴───────────────┐                       │
│            │  Source Fetchers              │                       │
│            │  ├── GitHub (CHANGELOG.md)    │                       │
│            │  ├── Docs (migration guides)  │                       │
│            │  └── Release notes            │                       │
│            └───────────────────────────────┘                       │
└─────────────────────────────────────────────────────────────────────┘
                          │
                          ▼
┌─────────────────────────────────────────────────────────────────────┐
│                    Migration Engine (Tiered)                        │
│  Tier 1: AST Transforms  │  Tier 2: KB-Guided  │  Tier 3: LLM      │
│  (deterministic)         │  (context + LLM)    │  (fallback)       │
└─────────────────────────────────────────────────────────────────────┘
  1. Fetch Knowledge: Discovers and fetches changelogs, migration guides from GitHub/PyPI
  2. Parse Changes: Uses LLM to extract breaking changes with confidence levels (HIGH/MEDIUM/LOW)
  3. Scan Codebase: Finds imports and usage of the target library
  4. Tiered Migration:
    • Tier 1: Deterministic AST transforms for known libraries (pydantic, fastapi, sqlalchemy, pandas, requests)
    • Tier 2: Knowledge base guided migration with LLM assistance
    • Tier 3: Pure LLM migration for unknown patterns
  5. Validate: Runs syntax checks and optionally your test suite
  6. Report: Shows a detailed diff with explanations for each change

Pydantic v1 → v2 Transforms

PyResolve handles the following Pydantic migrations automatically:

  • Config class → model_config = ConfigDict(...)
  • @validator@field_validator with @classmethod
  • @root_validator@model_validator
  • .dict().model_dump()
  • .json().model_dump_json()
  • .schema().model_json_schema()
  • .parse_obj().model_validate()
  • orm_mode = Truefrom_attributes = True
  • Field(regex=...)Field(pattern=...)

Configuration

PyResolve can be configured via pyproject.toml:

[tool.pyresolve]
# Path patterns to exclude from scanning (defaults: .pyresolve/*, tests/*, .venv/*, venv/*)
exclude = ["tests/*", "migrations/*"]

# Enable/disable LLM fallback
use_llm = true

# Anthropic API key (can also use ANTHROPIC_API_KEY env var)
# anthropic_api_key = "sk-..."

Pricing

PyResolve uses a tiered pricing model:

Tier Price Features
Free $0/month Tier 1 deterministic transforms (Pydantic, FastAPI, SQLAlchemy, Pandas, Requests)
Pro $19/month Tier 2 KB-guided LLM migrations for any library
Unlimited $49/month Tier 3 pure LLM migrations + priority support

How it works:

  • Tier 1 (Free): Runs entirely locally using deterministic AST transforms. No account required.
  • Tier 2/3 (Paid): LLM-powered migrations are processed through the PyResolve API to ensure quality and manage costs.
# Login to access Pro/Unlimited features
pyresolve login

# Check your current plan and usage
pyresolve quota

License

This software is licensed under the MIT License.

You are free to use, modify, and distribute this software. The CLI tool and all transforms are fully open source.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

codeshift-0.2.1.tar.gz (109.0 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

codeshift-0.2.1-py3-none-any.whl (125.4 kB view details)

Uploaded Python 3

File details

Details for the file codeshift-0.2.1.tar.gz.

File metadata

  • Download URL: codeshift-0.2.1.tar.gz
  • Upload date:
  • Size: 109.0 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for codeshift-0.2.1.tar.gz
Algorithm Hash digest
SHA256 ef3111cd2db98b24c761a47e43bd33541504ccea9b390ea704b51a0cf7157fba
MD5 721bb813a4e35c2168d9ba63d7fa25b1
BLAKE2b-256 3fbd17dab0082d22b518f7c46bf17dedf18834048ff08350736535b981acc854

See more details on using hashes here.

Provenance

The following attestation bundles were made for codeshift-0.2.1.tar.gz:

Publisher: publish.yml on Ragab-Technologies/PyResolve

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file codeshift-0.2.1-py3-none-any.whl.

File metadata

  • Download URL: codeshift-0.2.1-py3-none-any.whl
  • Upload date:
  • Size: 125.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for codeshift-0.2.1-py3-none-any.whl
Algorithm Hash digest
SHA256 a703ec41d73312ba837248cfca09d9717d7835fb0ba767d4999a5f747c19ff6b
MD5 dd1e887680363afc13e909516ed7391a
BLAKE2b-256 ada2ff95a6fc2a31c2c87bca48bd93ef95d5ecc3970635ebba947fbd49b54dbb

See more details on using hashes here.

Provenance

The following attestation bundles were made for codeshift-0.2.1-py3-none-any.whl:

Publisher: publish.yml on Ragab-Technologies/PyResolve

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page