Add preliminary Nexus support (#458)

* update

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* update

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* clarify

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update localBackendWorker.js

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

docs/app.md

@@ -9,9 +9,10 @@ Loading very large (>5M tips) trees in Taxonium works better in the desktop appl
 You can download the Taxonium desktop app for Windows, MacOS, or Linux from the [releases](https://github.com/theosanderson/taxonium/releases) page.
 
 :::{note}
-On MacOS you will need to "unquarantine" the app, because the code has not been signed. Before you do this, the application will claim that it is "damaged and can't be opened". 
+On MacOS you will need to "unquarantine" the app, because the code has not been signed. Before you do this, the application will claim that it is "damaged and can't be opened".
 
 Please use the terminal command below to achieve this:
+
 ```
 sudo xattr -r -d com.apple.quarantine "/Applications/Taxonium.app"
 ```

taxonium_electron/src/main.js

@@ -72,7 +72,11 @@ const setup = (mainWindow, args) => {
     stdio: ["pipe", "pipe", "pipe", "ipc"],
   });
 
-  console.log("Executing", binaryPath, [max_old_space_arg, scriptPath, ...args]);
+  console.log("Executing", binaryPath, [
+    max_old_space_arg,
+    scriptPath,
+    ...args,
+  ]);
   fork_id = p.pid;
 
   setTimeout(() => {

taxonium_web_client/src/components/InputSupplier.jsx

@@ -16,6 +16,7 @@ function formatBytes(bytes, decimals = 2) {
 const prettyTypes = {
   jsonl: "Taxonium JSONL",
   nwk: "Newick tree",
+  nexus: "Nexus tree (Nexus support is experimental)",
   meta_tsv: "Metadata TSV",
   meta_csv: "Metadata CSV",
   nextstrain: "Nextstrain JSON",

taxonium_web_client/src/hooks/useInputHelper.js

@@ -20,6 +20,11 @@ function guessType(file_object) {
   if (tree_extensions.includes(file_extension)) {
     return "nwk";
   }
+
+  if (file_extension === "nexus" || file_extension === "nex") {
+    return "nexus";
+  }
+
   if (file_extension === "jsonl") {
     return "jsonl";
   }
@@ -112,7 +117,11 @@ export const useInputHelper = ({
       return ["invalid", "You can only use a single metadata file"];
     }
     // can't have more than one tree file
-    if (inputs.filter((input) => input.filetype === "nwk").length > 1) {
+    if (
+      inputs.filter(
+        (input) => input.filetype === "nwk" || input.filetype === "nexus"
+      ).length > 1
+    ) {
       return ["invalid", "You can only use a single tree file"];
     }
     if (inputs.some((input) => input.filetype === "unknown")) {
@@ -121,7 +130,9 @@ export const useInputHelper = ({
     // must have a tree file or a jsonl
     if (
       inputs.filter((input) => input.filetype === "jsonl").length === 0 &&
-      inputs.filter((input) => input.filetype === "nwk").length === 0 &&
+      inputs.filter(
+        (input) => input.filetype === "nwk" || input.filetype === "nexus"
+      ).length === 0 &&
       inputs.filter((input) => input.filetype === "nextstrain").length === 0
     ) {
       return [
@@ -152,7 +163,10 @@ export const useInputHelper = ({
           input.filetype.startsWith("meta_")
         );
         const tree_file = inputs.find(
-          (input) => input.filetype === "nwk" || input.filetype === "nextstrain"
+          (input) =>
+            input.filetype === "nwk" ||
+            input.filetype === "nextstrain" ||
+            input.filetype === "nexus"
         );
         const newQuery = {
           treeUrl: tree_file.name,
@@ -192,7 +206,10 @@ export const useInputHelper = ({
 
       // if there is a tree file find it
       const tree_file = inputs.find(
-        (input) => input.filetype === "nwk" || input.filetype === "nextstrain"
+        (input) =>
+          input.filetype === "nwk" ||
+          input.filetype === "nextstrain" ||
+          input.filetype === "nexus"
       );
 
       upload_obj.filename = tree_file.name;

taxonium_web_client/src/utils/nexusToNewick.js

@@ -0,0 +1,43 @@
+// get nexusString from tree.nexus
+function nexusToNewick(nexusString) {
+  // get Translate section if present
+  const translateBlock = nexusString.match(/Translate(.*?);/gims);
+  const translate = translateBlock[0];
+
+  let translations = {};
+
+  // get all the translations
+  translate.split("\n").forEach((line) => {
+    line = line.trim();
+    const parts = line.split(" ");
+    if (parts.length === 2) {
+      const key = parts[0];
+      const value = parts[1].replace(",", "");
+      translations[key] = value;
+    }
+  });
+
+  // get all the trees
+
+  const treeBlock = nexusString.match(/tree TREE1 = (.*?;)/gims);
+
+  // get the Newick string from the tree block
+  const newickString = treeBlock[0].match(/\((.*?)\).+;/gims)[0];
+
+  //remove comments, which are indicated by [...]
+
+  const newick = newickString.replace(/\[(.*?)\]/gims, "");
+
+  // translate the taxon labels in the Newick string
+  const translatedNewickString = newick.replace(
+    /([^:\,\(\)]+)/gims,
+    (match) => {
+      //console.log(translations[match])
+      return translations[match] || match;
+    }
+  );
+
+  return translatedNewickString;
+}
+
+export default nexusToNewick;

taxonium_web_client/src/utils/processNewick.js

@@ -8,6 +8,7 @@ import {
 import pako from "pako";
 import axios from "axios";
 import reduceMaxOrMin from "./reduceMaxOrMin";
+import nexusToNewick from "../utils/nexusToNewick.js";
 const emptyList = [];
 
 function removeSquareBracketedComments(str) {
@@ -113,6 +114,11 @@ export async function processNewick(data, sendStatusMessage) {
 
   the_data = await fetch_or_extract(data, sendStatusMessage, "tree");
 
+  console.log("data.filetype", data.filetype);
+  if (data.filetype == "nexus") {
+    the_data = nexusToNewick(the_data);
+  }
+
   sendStatusMessage({
     message: "Parsing Newick file",
   });

taxonium_web_client/src/webworkers/localBackendWorker.js

@@ -213,7 +213,7 @@ onmessage = async (event) => {
     data.type === "upload" &&
     data.data &&
     data.data.filename &&
-    data.data.filetype === "nwk"
+    (data.data.filetype === "nwk" || data.data.filetype === "nexus")
   ) {
     console.log("got nwk file", data.data);
     data.data.useDistances = true;