cheerio-to-text 0.1.2 → 0.2.0

package/lib/esm/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,OAAO,EAAE,UAAU,EAAE,QAAQ,EAAW,OAAO,EAAE,MAAM,SAAS,CAAA;AAa9E,wBAAgB,MAAM,CACpB,IAAI,EAAE,UAAU,GAAG,QAAQ,GAAG,MAAM,GAAG,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,GAChE,MAAM,CA8BR"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,OAAO,EAAE,UAAU,EAAE,QAAQ,EAAW,OAAO,EAAE,MAAM,SAAS,CAAA;AAc9E,wBAAgB,MAAM,CACpB,IAAI,EAAE,UAAU,GAAG,QAAQ,GAAG,MAAM,GAAG,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,GAChE,MAAM,CAyCR"}

package/lib/esm/index.js

@@ -6,6 +6,7 @@ const inlineElements = new Set(`a,abbr,acronym,audio,b,bdi,bdo,big,br,button,can
   tt,u,var,video,wbr`
     .split(",")
     .map((s) => s.trim()));
+const isBlockTag = (tagName) => !inlineElements.has(tagName);
 export function render(node) {
     let root = null;
     if (typeof node === "string") {
@@ -23,24 +24,31 @@ export function render(node) {
     let text = "";
     function enter(element) {
         if (element.type === "text") {
-            if (element.data.trim())
-                text += element.data;
+            text += element.data;
         }
-        else if (element.type === "tag") {
-            if (!inlineElements.has(element.tagName)) {
-                if (text.at(-1) !== "\n")
-                    text += "\n";
+    }
+    function leave(element) {
+        if (element.type === "tag") {
+            // console.log({ LEAVING: element.type, tagName: element.tagName })
+            if (isBlockTag(element.tagName)) {
+                text += "\n";
             }
         }
     }
-    walk(root, enter);
-    return text.trim();
+    walk(root, enter, leave);
+    return text
+        .trim()
+        .split(/\n+/g)
+        .map((line) => line.trim())
+        .filter(Boolean)
+        .join("\n");
 }
-function walk(root, enter) {
+function walk(root, enter, leave) {
     enter(root);
     if (root.type === "tag") {
         for (const child of root.children) {
-            walk(child, enter);
+            walk(child, enter, leave);
         }
     }
+    leave(root);
 }

package/lib/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,OAAO,EAAE,UAAU,EAAE,QAAQ,EAAW,OAAO,EAAE,MAAM,SAAS,CAAA;AAa9E,wBAAgB,MAAM,CACpB,IAAI,EAAE,UAAU,GAAG,QAAQ,GAAG,MAAM,GAAG,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,GAChE,MAAM,CA8BR"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,OAAO,EAAE,UAAU,EAAE,QAAQ,EAAW,OAAO,EAAE,MAAM,SAAS,CAAA;AAc9E,wBAAgB,MAAM,CACpB,IAAI,EAAE,UAAU,GAAG,QAAQ,GAAG,MAAM,GAAG,OAAO,GAAG,OAAO,CAAC,OAAO,CAAC,GAChE,MAAM,CAyCR"}

package/lib/index.js

@@ -19,6 +19,7 @@ var cheerio_1 = __importDefault(require("cheerio"));
 var inlineElements = new Set("a,abbr,acronym,audio,b,bdi,bdo,big,br,button,canvas,cite,code,data,\n  datalist,del,dfn,em,embed,i,iframe,img,input,ins,kbd,label,map,mark,\n  meter,noscript,object,output,picture,progress,q,ruby,s,samp,script,\n  select,slot,small,span,strong,sub,sup,svg,template,textarea,time,\n  tt,u,var,video,wbr"
     .split(",")
     .map(function (s) { return s.trim(); }));
+var isBlockTag = function (tagName) { return !inlineElements.has(tagName); };
 function render(node) {
     var root = null;
     if (typeof node === "string") {
@@ -36,28 +37,34 @@ function render(node) {
     var text = "";
     function enter(element) {
         if (element.type === "text") {
-            if (element.data.trim())
-                text += element.data;
+            text += element.data;
         }
-        else if (element.type === "tag") {
-            if (!inlineElements.has(element.tagName)) {
-                if (text.at(-1) !== "\n")
-                    text += "\n";
+    }
+    function leave(element) {
+        if (element.type === "tag") {
+            // console.log({ LEAVING: element.type, tagName: element.tagName })
+            if (isBlockTag(element.tagName)) {
+                text += "\n";
             }
         }
     }
-    walk(root, enter);
-    return text.trim();
+    walk(root, enter, leave);
+    return text
+        .trim()
+        .split(/\n+/g)
+        .map(function (line) { return line.trim(); })
+        .filter(Boolean)
+        .join("\n");
 }
 exports.render = render;
-function walk(root, enter) {
+function walk(root, enter, leave) {
     var e_1, _a;
     enter(root);
     if (root.type === "tag") {
         try {
             for (var _b = __values(root.children), _c = _b.next(); !_c.done; _c = _b.next()) {
                 var child = _c.value;
-                walk(child, enter);
+                walk(child, enter, leave);
             }
         }
         catch (e_1_1) { e_1 = { error: e_1_1 }; }
@@ -68,4 +75,5 @@ function walk(root, enter) {
             finally { if (e_1) throw e_1.error; }
         }
     }
+    leave(root);
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "cheerio-to-text",
-  "version": "0.1.2",
+  "version": "0.2.0",
   "description": "Turn a Cheerio object into plain text",
   "repository": {
     "type": "git",