@ca-plant-list/ca-plant-list 0.4.14 → 0.4.16

package/scripts/inatobsphotos.js

@@ -0,0 +1,113 @@
+#!/usr/bin/env node
+
+import fs from "fs";
+import cliProgress from "cli-progress";
+import { stringify } from "csv-stringify";
+import path from "path";
+
+import { Program } from "../lib/program.js";
+import { Taxa } from "../lib/taxa.js";
+import { sleep } from "../lib/util.js";
+
+// While I'm guessing the products of this data will be non-commercial, it's
+// not clear how they'll be licensed so the ShareAlike clause is out, and
+// they'll probably be derivative works so the "No Derivatives" clause should
+// be respected.
+const ALLOWED_LICENSE_CODES = ["cc0", "cc-by", "cc-by-nc"];
+
+const DEFAULT_FILENAME = "inatobsphotos.csv";
+
+/**
+ * @param {Taxon} taxon
+ * @param {InatObsPhotosCommandLineOptions} options
+ * @return {Promise<InatApiObservation[]>}
+ */
+async function fetchObservationsForTaxon(taxon, options) {
+    const inatTaxonId = taxon.getINatID();
+    if (!inatTaxonId) return [];
+    let url =
+        `https://api.inaturalist.org/v2/observations/?taxon_id=${inatTaxonId}` +
+        "&photo_license=" +
+        ALLOWED_LICENSE_CODES.join(",") +
+        "&order=desc" +
+        "&order_by=votes" +
+        "&per_page=5" +
+        "&fields=(observation_photos:(photo:(url:!t,attribution:!t,license_code:!t)))";
+    if (typeof options.inatObsQuery === "string") {
+        url += `&${options.inatObsQuery}`;
+    }
+    const resp = await fetch(url);
+    if (!resp.ok) {
+        const error = await resp.text();
+        throw new Error(`Failed to fetch taxa from iNat: ${error}`);
+    }
+    const json = await resp.json();
+    return json.results;
+}
+
+/**
+ * @param {InatObsPhotosCommandLineOptions} options
+ */
+async function getObsPhotos(options) {
+    console.log("[inatobsphotos.js] options", options);
+
+    const taxa = await Taxa.loadTaxa(options);
+    const targetTaxa = taxa.getTaxonList();
+
+    const filename = path.join("data", options.filename || DEFAULT_FILENAME);
+    const writableStream = fs.createWriteStream(filename);
+    const columns = ["name", "id", "ext", "licenseCode", "attrName"];
+    const stringifier = stringify({ header: true, columns: columns });
+    stringifier.pipe(writableStream);
+    const prog = new cliProgress.SingleBar({
+        format: "Downloading [{bar}] {percentage}% | ETA: {eta}s | {value}/{total}",
+        etaBuffer: targetTaxa.length,
+    });
+    prog.setMaxListeners(100);
+    prog.start(targetTaxa.length, 0);
+
+    for (const taxon of targetTaxa) {
+        prog.increment();
+        const observations = await fetchObservationsForTaxon(taxon, options);
+        // Just get the CC-licensed ones, 5 per taxon should be fine (max is 20 on iNat). Whether or not
+        const photos = observations
+            .map((obs) => obs.observation_photos.map((op) => op.photo))
+            .flat()
+            .filter((photo) =>
+                ALLOWED_LICENSE_CODES.includes(photo.license_code),
+            )
+            .slice(0, 5);
+        for (const photo of photos) {
+            const row = [
+                taxon.getName(),
+                photo.id,
+                String(photo.url).split(".").at(-1),
+                // Need the license code to do attribution properly
+                photo.license_code,
+                // Photographers retain copyright for most CC licenses,
+                // except CC0, so attribution is a bit different
+                photo.attribution.match(/\(c\) (.*?),/)?.[1] ||
+                    photo.attribution.match(/uploaded by (.*)/)?.[1],
+            ];
+            stringifier.write(row);
+        }
+        await sleep(1_100);
+    }
+    prog.stop();
+}
+
+const program = Program.getProgram();
+program
+    .action(getObsPhotos)
+    .description("Write a CSV to datadir with iNaturalist observation photos")
+    .option(
+        "-q, --inat-obs-query <query>",
+        "Additional iNat observations API query terms to add, e.g. place_id=1234&d1=2020-01-01",
+    )
+    .option(
+        "-fn, --filename <filename>",
+        "Name of file to write to the data dir",
+        DEFAULT_FILENAME,
+    );
+
+await program.parseAsync();

package/scripts/inattaxonphotos.js

@@ -18,6 +18,8 @@ const ALLOWED_LICENSE_CODES = [
     "cc0", "cc-by", "cc-by-nc"
 ];
 
+const DEFAULT_FILENAME = "inattaxonphotos.csv";
+
 /**
  * @param {Taxon[]} taxa
  * @return {Promise<InatApiTaxon[]>}
@@ -35,7 +37,7 @@ async function fetchInatTaxa( taxa ) {
 }
 
 /**
- * @param {CommandLineOptions} options
+ * @param {InatTaxonPhotosCommandLineOptions} options
  */
 async function getTaxonPhotos( options ) {
     const errorLog = new ErrorLog(options.outputdir + "/errors.tsv");
@@ -46,7 +48,7 @@ async function getTaxonPhotos( options ) {
     );
     const targetTaxa = taxa.getTaxonList( );
 
-    const filename = path.join( "data", "inattaxonphotos.csv" );
+    const filename = path.join("data", options.filename || DEFAULT_FILENAME);
     const writableStream = fs.createWriteStream( filename );
     const columns = [
         "name",
@@ -80,7 +82,7 @@ async function getTaxonPhotos( options ) {
                 const row = [
                     taxon.getName(),
                     taxonPhoto.photo.id,
-                    taxonPhoto.photo.medium_url.split( "." ).at( -1 ),
+                    String( taxonPhoto.photo.medium_url ).split( "." ).at( -1 ),
                     // Need the license code to do attribution properly
                     taxonPhoto.photo.license_code,
                     // Photographers retain copyright for most CC licenses,
@@ -101,6 +103,10 @@ async function getTaxonPhotos( options ) {
 }
 
 const program = Program.getProgram();
-program.action(getTaxonPhotos).description( "Write a CSV to datadir with iNaturalist taxon photos" );
+program.action(getTaxonPhotos).description( "Write a CSV to datadir with iNaturalist taxon photos" )
+    .option(
+        "-fn, --filename <filename>",
+        "Name of file to write to the data dir"
+    );
 
 await program.parseAsync();

package/types/classes.d.ts

@@ -16,7 +16,7 @@ declare class Config {
 }
 
 declare class ErrorLog {
-    log(...args: string[]): void;
+    log(...args: any[]): void;
     write(): void;
 }
 
@@ -83,6 +83,7 @@ declare class Taxa {
     getFlowerColors(): FlowerColor[];
     getTaxon(name: string): Taxon;
     getTaxonList(): Taxon[];
+    hasSynonym(name: string): boolean;
     isSubset(): boolean;
 }
 
@@ -92,6 +93,7 @@ declare class TaxaCol {
     title: string;
 }
 
+type StatusCode = "N" | "NC" | "U" | "X";
 declare class Taxon {
     constructor(data: TaxonData, genera: Genera, meta: any);
     getBaseFileName(): string;
@@ -103,6 +105,7 @@ declare class Taxon {
     getCalscapeCommonName(): string | undefined;
     getCalscapeName(): string;
     getCESA(): string | undefined;
+    getCNDDBRank(): string | undefined;
     getCommonNames(): string[];
     getFamily(): Family;
     getFESA(): string | undefined;
@@ -110,21 +113,25 @@ declare class Taxon {
     getFlowerColors(): string[] | undefined;
     getGenus(): Genus;
     getGenusName(): string;
+    getGlobalRank(): string | undefined;
     getHTMLLink(
         href: boolean | string | undefined,
         includeRPI?: boolean,
     ): string;
     getINatID(): string;
     getINatName(): string;
+    getINatSyn(): string | undefined;
     getINatTaxonLink(): string;
     getJepsonID(): string;
     getLifeCycle(): string;
     getName(): string;
     getPhotos(): Photo[];
+    getRPIID(): string | undefined;
     getRPIRank(): string;
     getRPIRankAndThreat(): string;
     getRPIRankAndThreatTooltip(): string;
     getRPITaxonLink(): string;
+    getStatus(): StatusCode;
     getStatusDescription(config: Config): string;
     getSynonyms(): string[];
     isCANative(): boolean;
@@ -147,7 +154,7 @@ declare class TaxonData {
     life_cycle: string;
     "RPI ID": string;
     SRank: string;
-    status: string;
+    status: StatusCode;
     taxon_name: string;
 }
 
@@ -194,14 +201,32 @@ declare class InatCsvPhoto {
     attrName: string;
 }
 
+declare class InatApiPhoto {
+    id: number;
+    attribution: string;
+    license_code: InatLicenseCode;
+    medium_url?: string;
+    url?: string;
+}
+
 declare class InatApiTaxon {
     id: number;
     taxon_photos: {
-        photo: {
-            id: number;
-            attribution: string;
-            license_code: InatLicenseCode;
-            medium_url: string;
-        };
+        photo: InatApiPhoto;
     }[];
 }
+
+declare class InatApiObservation {
+    observation_photos: {
+        photo: InatApiPhoto;
+    }[]
+}
+
+declare class InatObsPhotosCommandLineOptions extends CommandLineOptions {
+    filename?: string;
+    inatObsQuery?: string;
+}
+
+declare class InatTaxonPhotosCommandLineOptions extends CommandLineOptions {
+    filename?: string;
+}

package/.github/workflows/main.yml

@@ -1,20 +0,0 @@
-on:
-    workflow_dispatch:
-    pull_request:
-    push:
-
-permissions:
-    contents: read
-
-jobs:
-    check:
-        runs-on: ubuntu-latest
-        steps:
-            - uses: actions/checkout@v4
-            - uses: actions/setup-node@v4
-              with:
-                  node-version: latest
-                  cache: "npm"
-            - run: npm update
-            - run: npx eslint
-            - run: npx tsc

package/.prettierrc

@@ -1,3 +0,0 @@
-{
-    "tabWidth": 4
-}

package/.vscode/settings.json

@@ -1,9 +0,0 @@
-{
-    "eslint.validate": ["javascript", "typescript"],
-    "json.schemas": [
-        {
-            "fileMatch": ["exceptions.json"],
-            "url": "./schemas/exceptions.schema.json"
-        }
-    ]
-}

package/eslint.config.mjs

@@ -1,13 +0,0 @@
-import globals from "globals";
-import pluginJs from "@eslint/js";
-
-/** @type {import('eslint').Linter.Config[]} */
-export default [
-    { files: ["**/*.{js,mjs,cjs}"], ignores: ["output/**"] },
-    {
-        languageOptions: {
-            globals: { ...globals.browser, ...globals.node, bootstrap: false },
-        },
-    },
-    pluginJs.configs.recommended,
-];

package/schemas/exceptions.schema.json

@@ -1,62 +0,0 @@
-{
-    "type": "object",
-    "additionalProperties": {
-        "type": "object",
-        "properties": {
-            "calflora": {
-                "type": "object",
-                "properties": {
-                    "badjepsonid": {
-                        "const": true
-                    },
-                    "native": { "type": "boolean" },
-                    "notintaxondata": {
-                        "const": true
-                    }
-                },
-                "additionalProperties": false
-            },
-            "calscape": {
-                "type": "object",
-                "properties": { "notnative": { "const": true } }
-            },
-            "comment": {
-                "type": "string"
-            },
-            "inat": {
-                "type": "object",
-                "properties": {
-                    "notintaxondata": {
-                        "const": true
-                    }
-                },
-                "additionalProperties": false
-            },
-            "jepson": {
-                "type": "object",
-                "properties": {
-                    "allowsynonym": {
-                        "const": true
-                    },
-                    "notineflora": {
-                        "const": true
-                    }
-                },
-                "additionalProperties": false
-            },
-            "rpi": {
-                "type": "object",
-                "properties": {
-                    "translation": {
-                        "type": "string"
-                    },
-                    "translation-to-rpi": {
-                        "type": "string"
-                    }
-                },
-                "additionalProperties": false
-            }
-        },
-        "additionalProperties": false
-    }
-}

package/tmp/config.json

@@ -1,21 +0,0 @@
-{
-    "calflora": {
-        "counties": [
-            "ALA",
-            "CCA"
-        ]
-    },
-    "inat": {
-        "project": "ebcnps"
-    },
-    "labels": {
-        "introduced": "Introduced to the East Bay",
-        "native": "Native to the East Bay",
-        "status-NC": "California native introduced to the East Bay"
-    },
-    "ebook": {
-        "filename": "ebplants",
-        "pub_id": "ebplants",
-        "title": "East Bay Plants"
-    }
-}

package/tmp/exceptions.json

@@ -1,93 +0,0 @@
-{
-    "Campanula exigua": {
-        "rpi": {
-            "translation-to-rpi": "Ravenella exigua"
-        }
-    },
-    "Campanula sharsmithiae": {
-        "rpi": {
-            "translation-to-rpi": "Ravenella sharsmithiae"
-        }
-    },
-    "Castilleja ambigua subsp. ambigua": {
-        "rpi": {
-            "translation-to-rpi": "Castilleja ambigua var. ambigua"
-        }
-    },
-    "Castilleja ambigua var. ambigua": {
-        "rpi": {
-            "translation": "Castilleja ambigua subsp. ambigua"
-        }
-    },
-    "Downingia ornatissima var. mirabilis": {
-        "inat": {
-            "notintaxondata": true
-        }
-    },
-    "Erysimum capitatum var. angustatum": {
-        "calflora": {
-            "badjepsonid": true
-        },
-        "jepson": {
-            "allowsynonym": true
-        }
-    },
-    "Heterotheca oregona var. rudis": {
-        "inat": {
-            "notintaxondata": true
-        }
-    },
-    "Horkelia californica var. frondosa": {
-        "inat": {
-            "notintaxondata": true
-        }
-    },
-    "Lupinus littoralis var. variicolor": {
-        "inat": {
-            "notintaxondata": true
-        }
-    },
-    "Malacothamnus hallii": {
-        "comment": "in CNPS Rare Plant Inventory",
-        "calflora": {
-            "notintaxondata": true
-        },
-        "jepson": {
-            "allowsynonym": true
-        },
-        "rpi": {
-            "translation": "Malacothamnus arcuatus var. elmeri"
-        }
-    },
-    "Malacothamnus arcuatus var. elmeri": {
-        "comment": "in CNPS Rare Plant Inventory as Malacothamnus hallii",
-        "rpi": {
-            "translation-to-rpi": "Malacothamnus hallii"
-        }
-    },
-    "Myosurus minimus subsp. apus": {
-        "comment": "in CNPS Rare Plant Inventory",
-        "jepson": {
-            "notineflora": true
-        }
-    },
-    "Ravenella exigua": {
-        "comment": "in CNPS Rare Plant Inventory as Ravenella",
-        "rpi": {
-            "translation": "Campanula exigua"
-        }
-    },
-    "Ravenella sharsmithiae": {
-        "rpi": {
-            "translation": "Campanula sharsmithiae"
-        }
-    },
-    "Streptanthus albidus subsp. peramoenus": {
-        "calflora": {
-            "notintaxondata": true
-        },
-        "jepson": {
-            "allowsynonym": true
-        }
-    }
-}