Polymer Processing Planner Local, offline planner for polymer processing. Stack:

Watch on YouTube

Install + Run

npm install
npm run dev

Open http://localhost:3000.

Project Structure

.
├─ src/
│  ├─ app.ts                  # Express app bootstrap
│  ├─ db.ts                   # SQLite connection + migrations
│  ├─ routes/                 # HTTP routes (auth, experiments, reports, notes, etc.)
│  ├─ services/               # Business logic (auth/report/tasks/qualification/markdown)
│  ├─ repos/                  # Data access layer (SQLite queries)
│  ├─ middleware/             # Auth/access/permission middlewares
│  ├─ domain/                 # Domain math/helpers (DOE imports/stats/designs)
│  ├─ views/                  # EJS pages + partials
│  ├─ public/                 # Frontend assets (app.css, app.js, illustrations)
│  └─ tests/                  # Integration tests
├─ dist/                      # Compiled output (`npm run build`)
├─ im_doe.sqlite              # Local SQLite database
├─ plan.md                    # Product/feature roadmap
└─ README.md

Auth Env (Stage 0 Prep)

Create a .env file based on .env.example and set:

• SESSION_SECRET
• ADMIN_EMAIL
• ADMIN_TEMP_PASSWORD
• GOOGLE_CLIENT_ID
• GOOGLE_CLIENT_SECRET

Authentication

• The first admin account is created on startup using ADMIN_EMAIL + ADMIN_TEMP_PASSWORD.
• ADMIN_TEMP_PASSWORD is used only for initial seeding. Changing it later does not update an existing admin password.
• After first login with the temp password, the admin must set a new password.
• Passwords are stored as bcrypt hashes (not in plain text).
• Roles: admin, manager, engineer, operator, viewer.
• Access:
  • admin / manager see all experiments and processes.
  • Process Owner can access and manage experiments in their process.
  • Experiment Owner can manage their experiment and sign/unsign report.
  • entity assignees get access to assigned experiment entities.

Core Flows

1. Open process list on /.
2. Open a specific process at /<process_route_code> (e.g. /injection).
3. Create an experiment in this process with recipes and machine assignment.
4. Run the 6‑step Scientific Molding qualification (each step has its own setup + runs).
5. Create multiple Detailed Optimization (DOE) studies under the same experiment.
6. Configure factors, generate runlists, and enter run data.
7. Review analysis (charts + heatmap + 3D when possible).
8. Generate and edit reports (multiple reports per experiment).
9. Track Tasks on a kanban board inside each experiment.

Planned: export data to CSV.

Routing Model

• Process list: /
• Process page: /<process_route_code> (configured in Process settings, admin-only)
• Experiment canonical URL: /<process_route_code>/<experiment_id>
• Legacy routes like /experiments/:id are still accepted and redirected/rewritten for compatibility.
• Process settings are available on the process page (Process settings dialog), not on the process cards.

Task Manager

• Tasks live inside each experiment and are shown as a 4-column kanban (Init / In progress / Done / Failed).
• Create a task with title, description, due date, and owner (defaults to the experiment owner).
• Link tasks to entities:
  • Qualification steps (1–6)
  • DOE studies
  • Reports (signature required)
• Task progress is calculated from linked entities; tasks without entities are driven by manual status moves.
• Reports can be signed inside the task popup (manager/engineer/admin).
• Each task has calendar actions: download .ics or open a Google Calendar link.

Entity Responsibility + Notifications

• Experiment owner (and admin/manager) can assign responsible users to:
  • Qualification steps
  • DOE studies
• Assignment creates/updates an automatic task for the assignee.
• Assignees get in-app notifications (with unread badge in top navigation).
• Profile page includes:
  • assigned entities list (entity + experiment links)
  • notifications feed with mark-read actions
• Report signature is restricted to the experiment owner.

Calendar (Current UX)

• Two calendar surfaces are available:
  • My Calendar on /me
  • Process Calendar on /<process_route_code> (collapsible panel)
• Calendar events include:
  • tasks (task)
  • DOE runs (run)
  • qualification runs (qual_run)
• Date updates:
  • drag one event to move one entity
  • move a selected group by dragging one selected event
  • bulk move selected events with Move selected + date input
• Selection model:
  • Shift/Cmd/Ctrl + click toggles entity selection
  • lasso selection (mouse rectangle) selects intersecting events
  • click empty calendar area or Clear selection to drop selection
• Event click opens a details popup with:
  • event type/status/date/owner
  • link to run/task
  • link to parent entity/experiment
• Process calendar remembers panel open/closed state in browser localStorage.

Process Model (Current)

• New DB entities:
  • process_types
  • processes (with owner_user_id, route_code, status)
  • experiments.process_id
• Startup migration ensures:
  • default type Injection
  • default type Compounding (Twin-Screw Extrusion)
  • default type Coating
  • default process Injection Default Process
  • default process Compounding Default Process
  • default process Coating Default Process
  • existing experiments are attached to default process

Notes & Lab Journal

• Notes are available in a bottom drawer on:
  • Experiment page
  • Qualification step page
  • DOE page
  • Run page
  • Report page
• Full note feed is available at /experiments/:id/journal.
• Notes are entity-linked (experiment, qualification_step, doe, run, report, task).
• Daily mode:
  • Ctrl/Cmd + Enter appends to your current daily note in the same entity context.
  • Ctrl/Cmd + Shift + Enter forces a new note.
• Filters:
  • text search
  • date filter
  • entity-only toggle (in entity drawers)
• Soft-delete is enabled for admin / manager.

Qualification Packs (by Process Type)

Qualification is process-specific (6-step pack is selected by process_type):

• Injection (Scientific Molding):

1. Rheology / Viscosity curve
2. Cavity balance
3. Pressure drop
4. Cosmetic process window
5. Gate seal study
6. Cooling time optimization

• Compounding (Twin-Screw Extrusion):

1. RTD / Residence Time Stability
2. SME Map / Energy Window
3. Melt Temperature / Thermal History Map
4. Feeding / Side-Feeder Qualification
5. Degassing / Moisture Control
6. Dispersion / Mixing Quality Check

• Coating (Water/Solvent/Extrusion Coatings):

1. Rheology Window
2. Wetting / Surface Energy Check
3. Coat Weight Calibration
4. Drying / Curing Window
5. Adhesion Qualification
6. Barrier / Functional Check

Implementation notes:

• Qualification step UI is process-specific:
  • Injection keeps Scientific Molding step-specific screens.
  • Compounding and Coating use an independent generic step editor (runs + fields), without cavity/rheology/gate-seal injection UI.
• Each qualification step is edited independently (/experiments/:id/qualification/:step): runs, values, assignee, and step fields are isolated per step.

DOE (Shared Engine, Process-Specific Defaults)

• DOE generation/analysis uses one shared module across process types.
• Defaults are process-specific:
  • active factors by process_type,
  • active measured outputs by process_type.
• Analysis reads analysis_run_values first, and falls back to run_values by field code when needed (useful for migration/demo data).

Reference book (Amazon search):

• Robust Process Development and Scientific Molding (Suhas Kulkarni): https://a.co/d/aDv52KL

Report Plan (Next Work)

The current report plan lives at report_plan in the project root.

Recent Changes (for handoff)

• Machine library now supports parameter tokens in the format %machineId:paramId%. These tokens can be used inside qualification setup inputs and custom fields.
• UI shows live previews for tokenized values; inputs keep the token, summaries display the resolved value.
• Step calculations resolve tokens at runtime (server + client), so values survive reloads.
• Machine edit page shows a small read-only token field next to each parameter for quick copy.
• Report generator now saves report configs per experiment (multiple reports per experiment).
• Report list lives inside the experiment, right after Detailed Optimization.
• Report editor (Editor.js) with seeded structure and embedded charts (rheology + process window).
• Editable report documents are stored in report_documents and opened via /reports/:id/editor.

Supported Recipe Import Formats

The importer accepts two common formats:

1) Matrix format

Component,Recipe A,Recipe B
Resin 1,50,60
Resin 2,50,40
Additive,3,2

• Column 0: component name
• Other columns: recipe name with PHR values

2) Two-row header (BPACKs style)

,Recipe A,,Recipe B,
,phr,,phr,
Resin 1,50,,60,
Resin 2,50,,40,
Additive,3,,2,

• Row 1 contains recipe names
• Row 2 contains phr under recipe columns
• Subsequent rows are components

Notes

• The SQLite database is im_doe.sqlite in this folder.
• Custom input/output fields are stored in the flexible param_definitions and run_values tables.
• SCREEN design is a sampled factorial (labeled in-app). For higher rigor, add a dedicated generator.

Scripts

• npm run dev - start with hot reload
• npm run build - compile to dist/
• npm run start - run compiled output