Skip to content

ROADMAP.md

Latest commit

 

History

History
210 lines (152 loc) · 7.69 KB

ROADMAP.md

File metadata and controls

210 lines (152 loc) · 7.69 KB
Raw

ScanFlow Roadmap

This document describes the planned evolution of ScanFlow beyond the current phase. Each phase has a clear goal, exit criteria, and concrete tasks.

The current state (post-refactor) covers:

  • New setp/getp API with SI units
  • Coarse approach + Z-limit + slider panels
  • Scan control panel with channel selector
  • Lock-in + I/V point spectroscopy
  • Cryo temperature readout
  • Drift detection + correction during automation
  • Overnight-safe recipes (DST suppression, configurable save folder)

Phase 2 — Live awareness (next, ~1 week)

Goal: the user can see what the STM is doing without leaving ScanFlow.

2.1 Live scan viewer

  • Embed a pyqtgraph.ImageView in a new Live View tab.
  • After each scan saves, load the .dat via createc.Createc_pyFile.DAT_IMG and update the viewer.
  • Add a channel selector (TOPOGRAPHY / CURRENT / DF / Lock-in X) and an auto-clip percentile slider.
  • Overlay the cumulative drift trail.

2.2 Watchdog + notifications

  • scanflow/notify/ module with three back-ends: sound, email, desktop.
  • Trigger on: scan finished, recipe finished, error, drift confidence < threshold.
  • Configurable per recipe.

2.3 Approach-status integration

  • Couple the approach result back to the Scan Control tab — auto-refresh parameters when the tip enters tunnelling.
  • Disable scan controls while approach is in progress.

Exit criteria: a user can run an overnight recipe, watch scans appear in real time, and get a phone notification when the run finishes or errors out.

Phase 3 — Spectroscopy maturity (~1–2 weeks)

Goal: ScanFlow can drive every spectroscopy mode the CreaTec supports, saving raw .VERT plus a sidecar JSON with metadata + lock-in state.

3.1 Multi-point and line spectroscopy

  • Wrap btn_vertspec_mult and btn_vertspec_line in dedicated panels.
  • Pick positions visually on the live image (click to add a marker).
  • Export marker lists as YAML, alongside the recipe.

3.2 Spectroscopy on a grid

  • Wrap btn_spectraongrid with a UI for defining the grid (origin, spacing, N×M).
  • Auto-name files: <datestamp>_grid_<i>_<j>.VERT.

3.3 Spectroscopy recipes

  • Extend MeasurementRecipe with a SpectroscopyStep type.
  • Allow mixed image+spectroscopy recipes (e.g. "scan, then grid-spec, then scan again").

3.4 dI/dV imaging

  • Add a panel that runs a normal scan with Lock-in X as a recorded channel, with the lock-in configured for bias modulation.

Exit criteria: a user can define a recipe that runs an overview scan, records dI/dV grid spectroscopy at picked points, and returns to image scanning — all unattended.

Phase 4 — AFM / qPlus support (~1 week)

Goal: the AFM Mode in stmafm.ini is fully usable from ScanFlow.

4.1 PLL / qPlus tuning panel

  • Wrap AFMController.find_resonance in a wizard:
    1. Broad scan → display amplitude vs frequency curve
    2. Auto-fit + zoom in
    3. Apply → set centre frequency, enable amplitude control
  • Tune controller bandwidth sliders.

4.2 Feedback channel switcher

  • A clearly labelled toggle between STM (current) and AFM (Δf) feedback, with safety prompts (Z-limit on/off, ramp setpoint slowly).

4.3 dF-Z spectroscopy

  • New spec mode in the spectroscopy panel: ramp Z while logging Δf.

Exit criteria: the existing manufacturer STM_AFM_operation.py and AFM_STM_operation.py example scripts can both be expressed entirely through ScanFlow.

Phase 5 — Sample mapping & navigation (~1–2 weeks)

Goal: track where the tip has been on the sample, and let the user revisit previous locations.

5.1 Sample map widget

  • A 2-D view of all scans taken in a session, plotted by their absolute offsets (slider position + scan-frame offset).
  • Click a scan → reload its parameters into the Scan Control tab.

5.2 Coordinate bookkeeping

  • Add an XYPosition accumulator that integrates slider pulses (with the user supplying a per-pulse nm calibration).
  • Save the position log to disk so it survives restarts.

5.3 "Return to previous location"

  • One-click reverse of slider motion to a prior bookmark.

Exit criteria: after moving across the sample for two hours, a user can visually identify and re-approach to any earlier scan area.

Phase 6 — Robustness & dev experience (~1 week)

Goal: ScanFlow can be developed and tested without the instrument, and gracefully handles real-world failures.

6.1 Mock STM

  • scanflow.core.mock.MockSTMClient that simulates the COM API.
  • Generates synthetic images (with controllable drift, noise, atomic lattice).
  • Used by tests and as an offline-mode toggle in the GUI.

6.2 Comprehensive tests

  • pytest-qt integration to test GUI panels with the mock client.
  • Property-based tests for the recipe builders.
  • Smoke test for every panel (boot, click around, no exceptions).

6.3 Error handling

  • Per-call retry policy for COM operations (transient errors are common).
  • Recipe-level "on error" handler: stop / pause / retry / continue.
  • Crash log with the last 1000 lines of the session log.

6.4 Settings & preferences

  • ~/.scanflow/config.yaml with all defaults user-tweakable.
  • Settings dialog in the GUI.

Exit criteria: the test suite covers every public method on every controller and every GUI panel, all using the mock client.

Phase 7 — Smarter drift correction (~2 weeks, research-y)

Goal: beat the current phase-cross-correlation approach on tricky surfaces.

7.1 Feature-based tracking

  • Optional ORB/SIFT feature matching path for highly textured surfaces.
  • Compare cross-correlation vs feature shift; pick the higher-confidence.

7.2 Anisotropic drift model

  • Fit drift rate per axis from the last N corrections.
  • Predict the next drift instead of always doing an alignment scan.
  • Skip alignment scans when prediction confidence is high.

7.3 Drift compensation via STM's internal mechanism

  • Wire Drift_X[A./sec] and Drift_Y[A./sec] into the GUI.
  • Let ScanFlow estimate these and push them to the instrument so the DSP does the correction inline — eliminates the need for alignment scans.

7.4 Atomic-resolution drift sub-pixel

  • When the lattice is resolved, use lattice-vector tracking for sub-pixel accuracy (interface with AiSurf for lattice extraction).

Exit criteria: drift correction works on bias values where features are weak, and overnight runs need 30–50% fewer alignment scans.

Phase 8 — Integration with the wider ScanFlow ecosystem (~1 week)

Goal: ScanFlow feeds clean data into the existing lab tools without manual file shuffling.

8.1 ProbeFlow handoff

  • Optional folder-watcher that exports finished .dat files plus metadata as ProbeFlow-compatible JSON sidecars.
  • Single "Open last scan in ProbeFlow" button.

8.2 SpmImageTycoon export

  • Optional batch-export to the .sxm format used by SpmImageTycoon.

8.3 AiSurf trigger

  • Per-scan checkbox: "auto-analyse lattice with AiSurf" — runs the analysis after each scan and writes the result alongside the .dat.

Exit criteria: a user can run an overnight session and wake up to scans already lattice-analysed, organised by ProbeFlow, with the data ready for review.

Cross-cutting concerns (continuous)

  • Documentation: every public method on a controller has a docstring; the README and ROADMAP are kept in sync with reality.
  • Logging: every COM call logs at DEBUG; every user-visible action logs at INFO. Daily rotating log file at ~/.scanflow/logs/.
  • Performance: keep the GUI responsive under 1-Hz scan completion rates. Use QThreads for any operation that can block.
  • Versioning: SemVer; the recipe YAML format gets a schema_version field before the first 1.0 release.