spikelab

# SpikeLab SpikeLab is a Python library for loading, analyzing, visualizing, and exporting neuronal spike train data from multi-electrode array (MEA) electrophysiology experiments. --- ## Session Start — Environment Detection Run this detection sequence once at the start of every session: 1. Check if `spikelab` is importable in the default `python`: ```bash python -c "import spikelab; print('OK')" 2>/dev/null ``` If this succeeds, use bare `python` for all commands. 2. If it fails, check for a `spikelab` conda environment: ```bash conda env list 2>/dev/null | grep spikelab ``` If found, prefix all commands with `conda run -n spikelab` for the rest of the session. 3. If neither works, **read `INSTALL.md`** in this skill directory for full installation instructions. Prefer the conda approach (Option A) unless the user specifies otherwise. Use the detected invocation method consistently for all Python/pip commands in the session. --- ## Delegation Once SpikeLab is available, delegate to the appropriate in-repo skill based on the user's request: **Spike sorting** — the user wants to sort raw recordings, configure sorting parameters, curate units, assess sorting quality, or troubleshoot sorting failures: ``` SpikeLab/src/spikelab/agent/skills/spikelab-spikesorter/SKILL.md ``` It covers: `sort_recording`, `sort_multistream`, curation, QC figures, and loading/inspecting sorted results. Outputs are saved as `sorted_spikedata_curated.pkl` files that the analysis implementer can load. **Explanatory and conceptual questions** — the user asks what an analysis does, how a method works, what a result means, wants to understand the underlying neuroscience, or asks about SpikeLab's API and data structures: ``` SpikeLab/src/spikelab/agent/skills/spikelab-educator/SKILL.md ``` It contains concept definitions, interpretation guidance, literature references, and cross-references to documentation, examples, and source code. **Library development** — the user wants to integrate analysis code into the library, promote a script's computations to reusable methods, add new functionality, write tests, or expose methods through MCP: ``` SpikeLab/src/spikelab/agent/skills/spikelab-developer/SKILL.md ``` It covers the full integration pipeline: auditing scripts against existing methods, replacing reimplementations, adding novel methods, writing tests (main usage + edge cases), MCP exposure, and PR submission. **Data analysis and visualization** — loading sorted data, running analyses, producing figures, writing scripts, and any ambiguous requests that are not sorting-related or development-related: ``` SpikeLab/src/spikelab/agent/skills/spikelab-analysis-implementer/SKILL.md ``` It contains the full analysis workflow: API orientation via repo maps, script structure, workspace usage, figure conventions, and reporting. It knows how to load sorting outputs from `sorted_spikedata_curated.pkl` files. --- ## Repo Maps The repo maps do not ship with the repository — they must be generated after installation and after every library update. Without them, the analysis-implementer, educator, developer, and spike-sorter skills cannot orient themselves to the library API and will produce unreliable results. Run the `spikelab-map-updater` skill to generate `REPO_MAP.md` and `REPO_MAP_DETAILED.md`: ``` SpikeLab/src/spikelab/agent/skills/spikelab-map-updater/SKILL.md ``` --- ## Quick Reference | What | Where | |---|---| | Library source | `SpikeLab/src/spikelab/` | | Repo maps (API docs) | `SpikeLab/src/spikelab/agent/skills/spikelab-map-updater/REPO_MAP.md`, `REPO_MAP_DETAILED.md` | | Map updater skill | `SpikeLab/src/spikelab/agent/skills/spikelab-map-updater/SKILL.md` | | In-repo spike sorter skill | `SpikeLab/src/spikelab/agent/skills/spikelab-spikesorter/SKILL.md` | | In-repo developer skill | `SpikeLab/src/spikelab/agent/skills/spikelab-developer/SKILL.md` | | In-repo analysis skill | `SpikeLab/src/spikelab/agent/skills/spikelab-analysis-implementer/SKILL.md` | | In-repo educator skill | `SpikeLab/src/spikelab/agent/skills/spikelab-educator/SKILL.md` | | Installation guide | `INSTALL.md` (this skill directory) |