Packages
Megatron Core provides GPU-optimized building blocks and parallelism strategies for training large transformer models at scale, including tensor parallelism, pipeline parallelism, and mixed precision support.
Not recommended for simple single-GPU training or inference-only use cases.
Megatron-FSDP implements Fully Sharded Data Parallelism (FSDP) in native PyTorch to distribute training and inference of large models across multiple GPUs with configurable memory-communication trade-offs.
Python client library for the Meilisearch API, enabling document indexing, full-text search with typo tolerance, and semantic search capabilities against a Meilisearch server.
Install it if you are already running a Meilisearch server and need a straightforward Python client for indexing and search.
Provides both async and sync Python clients for the Meilisearch search API, enabling document indexing and full-text search operations.
meld3 is an HTML/XML templating system that separates template markup from rendering logic using ElementTree-based element manipulation and namespace-qualified identifiers.
Install only if you are maintaining legacy code that already depends on it (e.g., Supervisor versions before 4.1.0).
Meltano is a declarative, code-first data integration platform that orchestrates ELT workflows, connecting 600+ APIs and databases to move, transform, and test data at scale.
A Singer target that loads data into Snowflake from any Singer tap, supporting append, upsert, and overwrite load methods with schema flattening and stream mapping.
Install it if you are using Singer taps and need a reliable Snowflake destination; it is the standard choice for this integration pattern.
Mem0 adds a persistent, searchable memory layer to AI agents and assistants, storing and retrieving user preferences, session state, and interaction history across conversations.
Analyzes binary size and memory footprint of embedded firmware by extracting symbol-level details from ELF files and linker scripts, with support for multiple architectures and source file mapping.
However, verify the actual license before use, as the metadata does not specify one.
Provides Python bindings for a high-performance distributed key-value cache engine designed for LLM and graph reasoning inference workloads, with multi-tier memory pooling and cross-device data access.
Splits text into semantic chunks at delimiters (periods, newlines, custom patterns) using SIMD acceleration for high-throughput processing.
Decorator-based function result caching with configurable TTL, multiple eviction algorithms (LRU, LFU, FIFO), and support for unhashable arguments.
Memori is a Python SDK that automatically captures and recalls structured memory from LLM interactions and agent conversations, persisting context across sessions without requiring code changes to existing LLM calls.
Profiles memory consumption line-by-line in Python functions and tracks overall process memory usage over time using the psutil module.
However, it is dormant with no active maintenance.
Identifies and provides access to RAM-based temporary file directories on Linux, allowing you to create temporary files in memory instead of on physical disk.
However, it is dormant (no releases since 2020); install it only if the stable API meets your needs and you do not expect active maintenance or new features.
MemPalace stores conversation history and project files as verbatim text and retrieves them via semantic search, with a structured index (wings, rooms, drawers) and pluggable backends including ChromaDB, Qdrant, Milvus, and pgvector.
Install it if you work with Claude Code, need local RAG, or want to preserve and search your conversation history.
Memray is a memory profiler for Python that traces every function call to track memory allocations across Python code, native extensions, and the interpreter itself, generating reports like flame graphs and tables to analyze memory usage.
Memsearch builds a searchable semantic memory index from markdown files, enabling AI coding agents to recall and reuse conversation context across sessions and platforms through vector search and hybrid retrieval.
Provides a Python API to query properties of chemical elements, ions, and isotopes from the periodic table, with support for physical units via pint and data export to pandas.
A Python client library for the Cisco Meraki Dashboard API that automatically handles rate limiting, retries, pagination, and async/sync HTTP operations across all current API endpoints.
Provides Python bindings to integrate Mercado Pago's payment APIs, enabling you to create orders, process payments, and manage transactions programmatically.
Install it if you need to integrate Mercado Pago payments into a Python application.
Mercantile converts between geographic coordinates, Web Mercator tiles (XYZ), and quadkeys, and provides utilities to navigate tile hierarchies and generate GeoJSON tile shapes.
Install it if you're working with tile-based maps or need to convert between geographic and tile coordinate systems.
Mercurial is a distributed version control system that tracks and manages changes to source code across multiple developers and repositories.
However, consider your team's existing tooling; install Mercurial only if you have a specific reason to use it.
Implements 3-way merge of text sequences, combining changes from two independent branches against a common base to produce a merged result with conflict markers.
Merges nested dictionaries recursively, combining multiple source dicts into a destination dict with configurable strategies for handling conflicting keys.
However, be aware it is dormant—no active maintenance since 2021—so do not expect bug fixes or new features.
A Python client library for accessing Merge APIs, which aggregate data from multiple HR, ATS, CRM, ticketing, accounting, and file storage platforms through a unified interface.
However, verify the license terms before use, and confirm that the specific API categories you need are fully implemented in this version.
Mergify CLI is a command-line tool for managing stacked pull requests, merge queues, CI test results, and scheduled freezes on GitHub, distributed as a self-contained binary with no runtime dependencies.
Programmatically generates MermaidJS diagram markup (flowcharts, sequence diagrams) as Python objects, avoiding manual string concatenation of diagram syntax.
However, dormancy since June 2024 and version 0.0.3 mean incomplete features and potential gaps; verify that flowchart and sequence diagram support covers your use…
mermaid-py provides a Python interface to the Mermaid diagram library, letting you create flowcharts, state diagrams, sequence diagrams, and other visualizations by writing Mermaid syntax in Python.
However, note that diagram type support is incomplete (several diagram types remain unimplemented), and reliance on mermaid.ink for rendering may be a concern in…
A Python library for controlling Meross IoT devices (smart plugs, bulbs, switches, hubs, and sensors) over the internet via HTTP and MQTT APIs.
Install it if you own Meross devices and need programmatic control; avoid it only if you lack a Meross account or require support for devices not listed in the tested…
meshcat is a Python interface to a browser-based 3D WebGL viewer that lets you remotely control and visualize 3D geometries, scenes, and transformations via a scene graph.
Reads and writes mesh files in dozens of formats (STL, VTK, Gmsh, ANSYS, Exodus, and others), converting smoothly between them for scientific and engineering workflows.
Install it if you work with mesh files across multiple tools or need programmatic mesh I/O in Python.
meshioplusplus reads and writes mesh files in dozens of formats (STL, VTK, Gmsh, CGNS, Exodus, and many others) and converts between them, with a C++20 core that provides zero-copy numpy integration at the I/O boundary.
A Python library and CLI for communicating with Meshtastic mesh radio devices, sending and receiving messages, and accessing device operations through a publish-subscribe event model.
Meson is a build system that reads declarative meson.build configuration files and generates build files to orchestrate compilation and testing of projects.
Install it if you need a build system for projects targeting Python 3.10 or newer.
meson-python is a PEP 517 build backend that uses the Meson build system to configure and compile Python packages, particularly those with native extension modules in C, C++, Cython, Fortran, Pythran, or Rust.
A Python client for MessageBird's REST API that lets you send SMS messages, manage voice calls, check account balance, and interact with other MessageBird services.
However, do not install if you require active support, regular API updates, or features added after mid-2024.
Metaflow is a framework for building, prototyping, and deploying AI and ML systems, handling the full lifecycle from local development through production orchestration with support for scaling across cloud compute resources.
Provides decorators for Metaflow workflows to save and resume checkpoints, manage models, and load external models from Hugging Face Hub, enabling long-running jobs to recover from failures without losing progress.
Provides Netflix-developed extensions to Metaflow, including an improved Conda environment decorator that mixes Conda and PyPI packages, debugging tools, and Hugging Face authentication support.
However, be aware that this is experimental functionality without the stability guarantees of core Metaflow; features may change or be removed if merged upstream.