Workflow docs - diagram and draft presentation (#54)

* chroma error imports have changed

* tweak path for docs action re-run

* publish action to run on PR (will this work)

* backtrack, publish action to push on main

* Added plankton_app_deployment.drawio

* Update plankton_app_deployment.drawio

* add the default material from quarto-revealjs-template

* outline in presentation form to complement demos

* quick screenshots for a walkthrough

* add image links to the just-in-time presentation

* fix titles and ORCID, add notes in README

docs/diagrams/could_be/plankton_app_deployment.drawio

@@ -0,0 +1,96 @@
+<mxfile host="app.diagrams.net" agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:132.0) Gecko/20100101 Firefox/132.0" version="24.8.3">
+  <diagram name="Page-1" id="uENUqXIsOQj-Sv-_rjTT">
+    <mxGraphModel dx="875" dy="465" grid="1" gridSize="10" guides="1" tooltips="1" connect="1" arrows="1" fold="1" page="1" pageScale="1" pageWidth="827" pageHeight="1169" math="0" shadow="0">
+      <root>
+        <mxCell id="0" />
+        <mxCell id="1" parent="0" />
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-1" value="&lt;div&gt;NAS &lt;br&gt;&lt;/div&gt;&lt;div&gt;S:\&lt;br&gt;&lt;/div&gt;" style="shape=cylinder3;whiteSpace=wrap;html=1;boundedLbl=1;backgroundOutline=1;size=15;" parent="1" vertex="1">
+          <mxGeometry x="120" y="250" width="60" height="80" as="geometry" />
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-2" value="&lt;div&gt;JASMIN&lt;/div&gt;&lt;div&gt;s3 storage&lt;br&gt;&lt;/div&gt;" style="ellipse;shape=cloud;whiteSpace=wrap;html=1;" parent="1" vertex="1">
+          <mxGeometry x="660" y="280" width="120" height="80" as="geometry" />
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-7" value="Task scheduler" style="rounded=1;whiteSpace=wrap;html=1;fillColor=#CCFFCC;" parent="1" vertex="1">
+          <mxGeometry x="310" y="80" width="120" height="60" as="geometry" />
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-9" value="" style="edgeStyle=orthogonalEdgeStyle;rounded=0;orthogonalLoop=1;jettySize=auto;html=1;" parent="1" source="HB7L-xhzdcljNeKlnPmA-7" target="HB7L-xhzdcljNeKlnPmA-8" edge="1">
+          <mxGeometry relative="1" as="geometry" />
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-11" value="Triggers" style="edgeLabel;html=1;align=center;verticalAlign=middle;resizable=0;points=[];" parent="HB7L-xhzdcljNeKlnPmA-9" vertex="1" connectable="0">
+          <mxGeometry x="0.1381" y="-1" relative="1" as="geometry">
+            <mxPoint as="offset" />
+          </mxGeometry>
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-15" style="edgeStyle=orthogonalEdgeStyle;rounded=0;orthogonalLoop=1;jettySize=auto;html=1;entryX=0.049;entryY=0.49;entryDx=0;entryDy=0;entryPerimeter=0;" parent="1" source="9WAs7ZXHk1DkFdnagY5z-1" target="HB7L-xhzdcljNeKlnPmA-2" edge="1">
+          <mxGeometry relative="1" as="geometry" />
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-16" value="&lt;div&gt;Write access&lt;/div&gt;" style="edgeLabel;html=1;align=center;verticalAlign=middle;resizable=0;points=[];" parent="HB7L-xhzdcljNeKlnPmA-15" vertex="1" connectable="0">
+          <mxGeometry x="-0.0696" relative="1" as="geometry">
+            <mxPoint as="offset" />
+          </mxGeometry>
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-19" value="" style="edgeStyle=orthogonalEdgeStyle;rounded=0;orthogonalLoop=1;jettySize=auto;html=1;" parent="1" source="HB7L-xhzdcljNeKlnPmA-8" target="HB7L-xhzdcljNeKlnPmA-18" edge="1">
+          <mxGeometry relative="1" as="geometry" />
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-20" value="Read/write access" style="edgeLabel;html=1;align=center;verticalAlign=middle;resizable=0;points=[];" parent="HB7L-xhzdcljNeKlnPmA-19" vertex="1" connectable="0">
+          <mxGeometry x="-0.2753" y="-2" relative="1" as="geometry">
+            <mxPoint as="offset" />
+          </mxGeometry>
+        </mxCell>
+        <mxCell id="9WAs7ZXHk1DkFdnagY5z-2" value="" style="edgeStyle=orthogonalEdgeStyle;rounded=0;orthogonalLoop=1;jettySize=auto;html=1;" edge="1" parent="1" source="HB7L-xhzdcljNeKlnPmA-8" target="9WAs7ZXHk1DkFdnagY5z-1">
+          <mxGeometry relative="1" as="geometry" />
+        </mxCell>
+        <mxCell id="9WAs7ZXHk1DkFdnagY5z-3" value="POSTS to " style="edgeLabel;html=1;align=center;verticalAlign=middle;resizable=0;points=[];" vertex="1" connectable="0" parent="9WAs7ZXHk1DkFdnagY5z-2">
+          <mxGeometry x="0.0278" y="-4" relative="1" as="geometry">
+            <mxPoint as="offset" />
+          </mxGeometry>
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-8" value="Task runner" style="rounded=1;whiteSpace=wrap;html=1;fillColor=#CCFFCC;" parent="1" vertex="1">
+          <mxGeometry x="310" y="250" width="120" height="60" as="geometry" />
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-12" value="&lt;h1 style=&quot;margin-top: 0px;&quot;&gt;Tasks&lt;/h1&gt;&lt;p&gt;* read TIFF images&lt;/p&gt;&lt;p&gt;* extract data and create metadata&lt;/p&gt;&lt;p&gt;* upload extracts to s3 storage&lt;br&gt;&lt;/p&gt;" style="text;html=1;whiteSpace=wrap;overflow=hidden;rounded=0;" parent="1" vertex="1">
+          <mxGeometry x="480" y="10" width="230" height="120" as="geometry" />
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-13" value="" style="endArrow=classic;html=1;rounded=0;exitX=0;exitY=0.5;exitDx=0;exitDy=0;entryX=0.971;entryY=0.423;entryDx=0;entryDy=0;entryPerimeter=0;" parent="1" source="HB7L-xhzdcljNeKlnPmA-8" target="HB7L-xhzdcljNeKlnPmA-1" edge="1">
+          <mxGeometry width="50" height="50" relative="1" as="geometry">
+            <mxPoint x="460" y="290" as="sourcePoint" />
+            <mxPoint x="510" y="240" as="targetPoint" />
+          </mxGeometry>
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-14" value="Read access" style="edgeLabel;html=1;align=center;verticalAlign=middle;resizable=0;points=[];" parent="HB7L-xhzdcljNeKlnPmA-13" vertex="1" connectable="0">
+          <mxGeometry x="-0.0066" relative="1" as="geometry">
+            <mxPoint as="offset" />
+          </mxGeometry>
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-18" value="Metadata storage" style="shape=cylinder3;whiteSpace=wrap;html=1;boundedLbl=1;backgroundOutline=1;size=15;fillColor=#CCFFCC;" parent="1" vertex="1">
+          <mxGeometry x="420" y="390" width="60" height="80" as="geometry" />
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-22" style="edgeStyle=orthogonalEdgeStyle;rounded=0;orthogonalLoop=1;jettySize=auto;html=1;entryX=1;entryY=0.5;entryDx=0;entryDy=0;entryPerimeter=0;" parent="1" source="HB7L-xhzdcljNeKlnPmA-21" target="HB7L-xhzdcljNeKlnPmA-18" edge="1">
+          <mxGeometry relative="1" as="geometry">
+            <mxPoint x="500" y="450" as="targetPoint" />
+            <Array as="points" />
+          </mxGeometry>
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-24" value="Read access" style="edgeLabel;html=1;align=center;verticalAlign=middle;resizable=0;points=[];" parent="HB7L-xhzdcljNeKlnPmA-22" vertex="1" connectable="0">
+          <mxGeometry x="-0.0122" relative="1" as="geometry">
+            <mxPoint as="offset" />
+          </mxGeometry>
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-25" value="" style="edgeStyle=orthogonalEdgeStyle;rounded=0;orthogonalLoop=1;jettySize=auto;html=1;" parent="1" source="HB7L-xhzdcljNeKlnPmA-21" target="HB7L-xhzdcljNeKlnPmA-2" edge="1">
+          <mxGeometry relative="1" as="geometry" />
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-26" value="Read access" style="edgeLabel;html=1;align=center;verticalAlign=middle;resizable=0;points=[];" parent="HB7L-xhzdcljNeKlnPmA-25" vertex="1" connectable="0">
+          <mxGeometry x="-0.0135" y="2" relative="1" as="geometry">
+            <mxPoint as="offset" />
+          </mxGeometry>
+        </mxCell>
+        <mxCell id="HB7L-xhzdcljNeKlnPmA-21" value="&lt;div&gt;Posit / Datalabs&lt;/div&gt;&lt;div&gt;(streamlit / fastapi)&lt;br&gt;&lt;/div&gt;" style="ellipse;shape=cloud;whiteSpace=wrap;html=1;" parent="1" vertex="1">
+          <mxGeometry x="630" y="420" width="120" height="70" as="geometry" />
+        </mxCell>
+        <mxCell id="9WAs7ZXHk1DkFdnagY5z-1" value="&lt;div&gt;Object store API &lt;br&gt;&lt;/div&gt;&lt;div&gt;(python / fastapi) &lt;br&gt;&lt;/div&gt;" style="rounded=1;whiteSpace=wrap;html=1;fillColor=#CCFFCC;" vertex="1" parent="1">
+          <mxGeometry x="510" y="150" width="130" height="40" as="geometry" />
+        </mxCell>
+      </root>
+    </mxGraphModel>
+  </diagram>
+</mxfile>

presentation/README.md

@@ -0,0 +1,7 @@
+# Overview presentation
+
+A developing overview of the aims and state of this project.
+
+Please see the seed project at [quarto-revealjs-template](https://github.com/ukceh-rse/quarto-revealjs-template) for detail documentation.
+
+In short, within this directory `quarto preview presentation.qmd`

presentation/images/.gitignore

@@ -0,0 +1 @@
+*.svg

presentation/presentation.qmd

@@ -0,0 +1,94 @@
+---
+title: Plankton ML Pipelines
+subtitle: Workflows for re-use of image machine learning models
+date: last-modified
+date-format: long
+
+format:
+  revealjs:
+    theme: [default, ukceh.scss]
+    logo: images/logo.png
+    footer: Slides published at <a href="https://ukceh-rse.github.io/quarto-revealjs-template">ukceh-rse.github.io/quarto-revealjs-template</a>
+    slide-number: c/t
+    show-slide-number: all
+    progress: true
+    embed-resources: true
+    title-slide-attributes:
+      data-background-image: images/PP_wave-01.png
+    header-includes: |
+      <script src="ukceh.js" type="application/javascript"></script>
+
+affiliations:
+  - id: ukceh
+    name: UK Centre for Ecology & Hydrology
+    url: https://ceh.ac.uk
+
+authors:
+  - name: Jo Walsh, Alba Gomez Segura
+    affiliation:
+      - ref: ukceh
+    email: [email protected]
+    url: https://github.com/metazool
+    orcid: 0000-0003-0268-2263
+
+bibliography: references.bib
+---
+
+## Plankton machine learning pipelines
+
+* Building on work done as a Turing collaboration
+* Intended to build reusable workflows for other image domains
+* APIs to simplify working with models and storage
+* Pipelines, classification and embeddings
+
+## Workflow - data, processing, cloud storage
+
+![workflow, high-level](images/workflow.png)
+
+## Luigi pipeline
+
+* Simple python package similar to R's targets, originally from Spotify
+* Usage from AMI and other Biodiversity projects
+* Great for prototyping, probably less so for production
+
+## Object store API 
+
+https://github.com/NERC-CEH/object_store_api/
+![](images/API_list_example.png)
+
+## Choice of models 
+
+* Several off-the-shelf models published by Turing Inst
+* One trained on CEFAS data several years ago, published through sci.vision
+* Several newer ones, oriented to marine plankton, both light and heavy to run
+
+https://github.com/alan-turing-institute/ViT-LASNet
+
+## Image embeddings
+
+* Image machine learning models are useful as a source of _embeddings_ 
+* E.g. a vector of numbers which the model uses as input to classification
+* Use of embeddings for search and retrieval is a common technique for working with Large Language Models
+
+## Embeddings visualiser
+
+![](images/embeddings_similarity_screenshot.png)
+* similarity search
+* clustering (simplest K-means clustering shown)
+
+## Image annotation
+
+https://github.com/NERC-CEH/cyto-ML - Ezra's original work with contributions from Phil and Erica
+
+Discussion about whether we can adapt an off-the-shelf package rather than DIY
+
+"Seeding success" project is using Label Studio which looks great
+
+## Thanks for listening! { .no-logo .no-footer .closing-slide .center background-image="images/PP_wave-02.png"}
+
+https://github.com/NERC-CEH/plankton_ml
+
+https://github.com/NERC-CEH/cyto-ML
+
+
+## References

presentation/ukceh.js

@@ -0,0 +1,49 @@
+
+function toggleLogo(event) {
+    var elements = document.querySelectorAll(".slide-logo");
+    if (event.currentSlide.matches('#title-slide') || event.currentSlide.classList.contains('no-logo')) {
+        // Hide the logo on the title slide
+        elements.forEach(function(elem) {
+            elem.style.display = "none";
+        });
+    } else {
+        // Show and position the logo on other slides
+        elements.forEach(function(elem) {
+            elem.style.display = "block";
+        });
+    }
+}
+
+function toggleFooter(event) {
+    var elements = document.querySelectorAll("div.footer-default");
+    if (event.currentSlide.matches('#title-slide') || event.currentSlide.classList.contains('no-footer')) {
+        // Hide the logo on the title slide
+        elements.forEach(function(elem) {
+            elem.style.display = "none";
+        });
+    } else {
+        // Show and position the logo on other slides
+        elements.forEach(function(elem) {
+            elem.style.display = "block";
+        });
+    }
+}
+
+window.addEventListener("load", (event) => {
+    var elements = document.querySelectorAll(".slide-logo");
+    elements.forEach(function(elem) {
+        elem.style.display = "none"; // Hide initially
+    });
+
+    var elements = document.querySelectorAll("div.footer-default");
+    elements.forEach(function(elem) {
+        elem.style.display = "none"; // Hide initially
+    });
+
+    Reveal.on("slidechanged", function(event) {
+        toggleLogo(event);
+        toggleFooter(event);
+    });
+});
+
+

presentation/ukceh.scss

@@ -0,0 +1,58 @@
+/*-- scss:defaults --*/
+
+$body-bg: whitesmoke !default;
+$body-color: black !default;
+$link-color: green !default;
+$presentation-heading-color: steelblue !default;
+$presentation-h1-font-size: 2.0em !default;
+$presentation-title-slide-text-align: left !default;
+
+/*-- scss:rules --*/
+
+
+.reveal .slide-logo {
+    width: auto !important;
+    height: auto !important;
+    max-height: 4rem !important;
+    left: 16px;
+}
+
+.reveal .slide-menu-button {
+    left: 8px;
+    bottom: unset;
+    top: 8px;
+}
+
+.reveal .slide-number {
+    top: unset !important;
+    right: 8px !important;
+    bottom: 8px !important;
+    font-size: 24px !important;
+}
+
+.reveal .slides > section:first-child {
+    color: white !important;
+}
+
+.reveal .slides > section:first-child h1 {
+    color: white !important;
+}
+
+.reveal .slide-number > a[href*='title-slide'] {
+  display: none;
+}
+
+.closing-slide {
+    color: white;
+}
+
+.closing-slide h1,
+.closing-slide h2,
+.closing-slide h3,
+.closing-slide h4,
+.closing-slide h5,
+.closing-slide h6 {
+    color: white;
+}
+
+

scripts/params.yaml

@@ -1,4 +1,4 @@
 cluster:
   n_clusters: 5
 
-collection: test-upload-alba
+collection: untagged-images-lana