@sillsdev/docu-notion 1.0.0-alpha.5 → 1.0.0-alpha.7

package/README.md

@@ -58,7 +58,7 @@ means that the ID is `0456aa5842946PRETEND4f37c97a0e5`.
 Try it out:
 
 ```
-npx @sillsdev/docu-notion -n secret_PRETEND123456789PRETEND123456789PRETEND6789 -r 0456aa5842946PRETEND4f37c97a0e5"
+npx @sillsdev/docu-notion -n secret_PRETEND123456789PRETEND123456789PRETEND6789 -r 0456aa5842946PRETEND4f37c97a0e5
 ```
 
 Likely, you will want to store these codes in your environment variables and then use them like this:

package/dist/plugins/embedTweaks.js

@@ -6,8 +6,26 @@ exports.gifEmbed = {
     regexMarkdownModifications: [
         {
             // I once saw a gif coming from Notion that wasn't a full
-            // url, which wouldn't work, hence the "http" requirement
-            regex: /\[.*?\]\((http.*?(\.(gif|GIF)))\)/,
+            // url, which wouldn't work, hence the "http" requirement.
+            //
+            // We only embed when the link's text is an AUTO-GENERATED label, not
+            // something the author typed. By the time this regex runs the label can
+            // take three auto shapes:
+            //   - empty `[]` (an embed)
+            //   - the literal `[bookmark]` placeholder notion-to-md emits for
+            //     bookmark blocks (seen when standardExternalLinkConversion isn't in
+            //     the config, e.g. the unit tests)
+            //   - the URL repeated as the label, `[http...gif](http...gif)`, which is
+            //     what standardExternalLinkConversion rewrites `[bookmark]` into in
+            //     the production config (it runs in an earlier phase, see transform.ts)
+            // A link the author gave real text, e.g. `[see animation](...gif)`, is
+            // none of these, so it's left as a clickable link rather than turned into
+            // an image (which would also throw the text away).
+            //
+            // The optional leading "!" lets us match (and replace) a link that
+            // another mod has already turned into an image, instead of prepending
+            // a second "!" and producing "!![](...)".
+            regex: /!?\[(?:bookmark|https?:\/\/[^\]]*\.(?:gif|GIF)[^\]]*)?\]\((http.*?(\.(gif|GIF)))\)/,
             replacementPattern: `![]($1)`,
         },
     ],
@@ -16,7 +34,14 @@ exports.imgurGifEmbed = {
     name: "imgur",
     regexMarkdownModifications: [
         {
-            regex: /\[.*?\]\((.*?imgur\.com\/.*?)\)/, // imgur.com
+            // Only embed links whose text is an auto-generated label (empty `[]`,
+            // `[bookmark]`, or the URL repeated as the label `[http...imgur.com/...]`),
+            // so that a link the author gave real text to is kept as a clickable link
+            // rather than converted to an image. See the longer note in gifEmbed above
+            // for why all three label shapes can occur here.
+            // The optional leading "!" lets us match a link that's already an image
+            // without prepending a second "!".
+            regex: /!?\[(?:bookmark|https?:\/\/[^\]]*imgur\.com\/[^\]]*)?\]\((.*?imgur\.com\/.*?)\)/, // imgur.com
             // imgur links to gifs need a .gif at the end, but the url they give you doesn't have one.
             replacementPattern: `![]($1.gif)`,
         },

package/dist/plugins/embedTweaks.spec.js

@@ -8,10 +8,38 @@ var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, ge
         step((generator = generator.apply(thisArg, _arguments || [])).next());
     });
 };
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
 Object.defineProperty(exports, "__esModule", { value: true });
 const log_1 = require("../log");
 const pluginTestRun_1 = require("./pluginTestRun");
 const embedTweaks_1 = require("./embedTweaks");
+const externalLinks_1 = require("./externalLinks");
+const default_docunotion_config_1 = __importDefault(require("../config/default.docunotion.config"));
+function paragraph(text) {
+    return {
+        type: "paragraph",
+        paragraph: {
+            rich_text: [
+                {
+                    type: "text",
+                    text: { content: text, link: null },
+                    annotations: {
+                        bold: false,
+                        italic: false,
+                        strikethrough: false,
+                        underline: false,
+                        code: false,
+                        color: "default",
+                    },
+                    plain_text: text,
+                    href: null,
+                },
+            ],
+        },
+    };
+}
 test("imgur", () => __awaiter(void 0, void 0, void 0, function* () {
     (0, log_1.setLogLevel)("verbose");
     const config = { plugins: [embedTweaks_1.imgurGifEmbed] };
@@ -25,6 +53,93 @@ test("imgur", () => __awaiter(void 0, void 0, void 0, function* () {
     ]);
     expect(result.trim()).toBe(`![](https://imgur.com/gallery/U8TTNuI.gif)`);
 }));
+// Regression test for the "!![](...)" bug: with both plugins active (the
+// default config order), the imgur mod turns a bare imgur link into an image,
+// then the gif mod sees that ".gif" image and must not prepend a second "!".
+test("imgur + gif together produce a single leading bang", () => __awaiter(void 0, void 0, void 0, function* () {
+    (0, log_1.setLogLevel)("verbose");
+    const config = { plugins: [embedTweaks_1.imgurGifEmbed, embedTweaks_1.gifEmbed] };
+    const result = yield (0, pluginTestRun_1.blocksToMarkdown)(config, [
+        {
+            object: "block",
+            id: "e36710d8-98ad-40dc-b41b-b376ebdd6894",
+            type: "bookmark",
+            bookmark: { caption: [], url: "https://imgur.com/E83qLj6" },
+        },
+    ]);
+    expect(result.trim()).toBe(`![](https://imgur.com/E83qLj6.gif)`);
+    expect(result).not.toContain("!![]");
+}));
+// An inline link the user gave real text to must stay a clickable link: we must
+// not turn it into an image (which also discards the text). Uses the default
+// config order so both mods get a crack at it.
+test("imgur link with author text is left as a clickable link", () => __awaiter(void 0, void 0, void 0, function* () {
+    (0, log_1.setLogLevel)("verbose");
+    const config = { plugins: [embedTweaks_1.imgurGifEmbed, embedTweaks_1.gifEmbed] };
+    const result = yield (0, pluginTestRun_1.blocksToMarkdown)(config, [
+        paragraph("all at once ([see animation](https://imgur.com/gcrxl5k))."),
+        paragraph("(See an animation of these [new overlay features](https://imgur.com/E83qLj6))"),
+    ]);
+    expect(result).toContain("[see animation](https://imgur.com/gcrxl5k)");
+    expect(result).toContain("[new overlay features](https://imgur.com/E83qLj6)");
+    // nothing should have been turned into an image
+    expect(result).not.toContain("![]");
+    expect(result).not.toContain(".gif");
+}));
+// Production-realistic: standardExternalLinkConversion runs in an earlier phase
+// and rewrites a Notion `[bookmark](url)` into `[url](url)` before the embed
+// regexes run. The embed must still recognize that as an auto-label and embed it.
+test("imgur bookmark still embeds after external-link conversion", () => __awaiter(void 0, void 0, void 0, function* () {
+    (0, log_1.setLogLevel)("verbose");
+    const config = {
+        plugins: [externalLinks_1.standardExternalLinkConversion, embedTweaks_1.imgurGifEmbed, embedTweaks_1.gifEmbed],
+    };
+    const result = yield (0, pluginTestRun_1.blocksToMarkdown)(config, [
+        {
+            object: "block",
+            id: "e36710d8-98ad-40dc-b41b-b376ebdd6894",
+            type: "bookmark",
+            bookmark: { caption: [], url: "https://imgur.com/gallery/U8TTNuI" },
+        },
+    ]);
+    expect(result.trim()).toBe(`![](https://imgur.com/gallery/U8TTNuI.gif)`);
+    expect(result).not.toContain("!![]");
+}));
+// ...but an authored inline link must survive that same pipeline as a link.
+test("authored imgur link survives external-link conversion as a link", () => __awaiter(void 0, void 0, void 0, function* () {
+    (0, log_1.setLogLevel)("verbose");
+    const config = {
+        plugins: [externalLinks_1.standardExternalLinkConversion, embedTweaks_1.imgurGifEmbed, embedTweaks_1.gifEmbed],
+    };
+    const result = yield (0, pluginTestRun_1.blocksToMarkdown)(config, [
+        paragraph("all at once ([see animation](https://imgur.com/gcrxl5k))."),
+    ]);
+    expect(result).toContain("[see animation](https://imgur.com/gcrxl5k)");
+    expect(result).not.toContain("![]");
+    expect(result).not.toContain(".gif");
+}));
+// Strongest guard: run the ACTUAL default production config (full plugin set
+// and order) rather than a hand-picked subset, so future config drift (plugin
+// reordering, the external-link converter changing) can't silently reopen the
+// bug. A Notion bookmark to imgur must embed; an authored inline link must stay
+// a clickable link.
+test("default production config: bookmark embeds, authored link stays a link", () => __awaiter(void 0, void 0, void 0, function* () {
+    (0, log_1.setLogLevel)("verbose");
+    const result = yield (0, pluginTestRun_1.blocksToMarkdown)(default_docunotion_config_1.default, [
+        {
+            object: "block",
+            id: "e36710d8-98ad-40dc-b41b-b376ebdd6894",
+            type: "bookmark",
+            bookmark: { caption: [], url: "https://imgur.com/gallery/U8TTNuI" },
+        },
+        paragraph("all at once ([see animation](https://imgur.com/gcrxl5k))."),
+    ]);
+    // the bookmark became an embedded gif...
+    expect(result).toContain("![](https://imgur.com/gallery/U8TTNuI.gif)");
+    expect(result).not.toContain("!![]");
+    // ...while the authored link is untouched
+    expect(result).toContain("[see animation](https://imgur.com/gcrxl5k)");
+}));
 test("gif", () => __awaiter(void 0, void 0, void 0, function* () {
     (0, log_1.setLogLevel)("verbose");
     const config = { plugins: [embedTweaks_1.gifEmbed] };

package/dist/plugins/internalLinks.js

@@ -29,7 +29,10 @@ function normalizeLinkBaseId(baseLinkId) {
         }
     }
     const withoutLeadingSlash = withoutQuery.replace(/^\/+/, "");
-    return withoutLeadingSlash;
+    // Newer Notion links can also use a relative "/p/<page-id>" form.
+    return withoutLeadingSlash.startsWith("p/")
+        ? withoutLeadingSlash.substring(2)
+        : withoutLeadingSlash;
 }
 // converts a url to a local link, if it is a link to a page in the Notion site
 // only here for plugins, notion won't normally be giving us raw urls (at least not that I've noticed)
@@ -125,7 +128,7 @@ exports.standardInternalLinkConversion = {
         // "Mention" links come in as full URLs, e.g. [link_to_page](https://www.notion.so/62f1187010214b0883711a1abb277d31)
         // Newer Notion links can also use app.notion.com, including /p/<page-id> URLs.
         // YOu can create them either with @+the name of a page, or by pasting a URL and then selecting the "Mention" option.
-        match: /\[([^\]]+)?\]\((?!mailto:)(https?:\/\/(?:www\.)?notion\.so\/[^),^/]+|https?:\/\/app\.notion\.com\/(?:p\/)?[^),^/]+|\/?[^),^/]+)\)/,
+        match: /\[([^\]]+)?\]\((?!mailto:)(https?:\/\/(?:www\.)?notion\.so\/[^),^/]+|https?:\/\/app\.notion\.com\/(?:p\/)?[^),^/]+|\/?(?:p\/)?[^),^/]+)\)/,
         convert: convertInternalLink,
     },
 };

package/dist/plugins/internalLinks.spec.js

@@ -121,6 +121,16 @@ test("parseLinkId extracts the page id from app.notion.com URLs", () => {
         fragmentId: "",
     });
 });
+test("parseLinkId extracts the page id from relative /p/ links", () => {
+    expect((0, internalLinks_1.parseLinkId)("/p/123456781234123412341234567890ab")).toEqual({
+        baseLinkId: "123456781234123412341234567890ab",
+        fragmentId: "",
+    });
+    expect((0, internalLinks_1.parseLinkId)("/p/123456781234123412341234567890ab#heading")).toEqual({
+        baseLinkId: "123456781234123412341234567890ab",
+        fragmentId: "#heading",
+    });
+});
 test("link to an existing page on this site that has no slug", () => __awaiter(void 0, void 0, void 0, function* () {
     const targetPageId = "123";
     const targetPage = (0, pluginTestRun_1.makeSamplePageObject)({
@@ -183,6 +193,68 @@ test("link to an existing page on this site that has no slug", () => __awaiter(v
     }, targetPage);
     expect(results.trim()).toBe(`Inline [great page](/${targetPageId}) the end.`);
 }));
+test("inline link using the newer relative /p/ form", () => __awaiter(void 0, void 0, void 0, function* () {
+    const targetPageId = "123";
+    const targetPage = (0, pluginTestRun_1.makeSamplePageObject)({
+        slug: undefined,
+        name: "Hello World",
+        id: targetPageId,
+    });
+    const results = yield getMarkdown({
+        type: "paragraph",
+        paragraph: {
+            rich_text: [
+                {
+                    type: "text",
+                    text: { content: "Inline ", link: null },
+                    annotations: {
+                        bold: false,
+                        italic: false,
+                        strikethrough: false,
+                        underline: false,
+                        code: false,
+                        color: "default",
+                    },
+                    plain_text: "Inline ",
+                    href: null,
+                },
+                {
+                    type: "text",
+                    text: {
+                        content: "great page",
+                        link: { url: `/p/${targetPageId}` },
+                    },
+                    annotations: {
+                        bold: false,
+                        italic: false,
+                        strikethrough: false,
+                        underline: false,
+                        code: false,
+                        color: "default",
+                    },
+                    plain_text: "great page",
+                    href: `/p/${targetPageId}`,
+                },
+                {
+                    type: "text",
+                    text: { content: " the end.", link: null },
+                    annotations: {
+                        bold: false,
+                        italic: false,
+                        strikethrough: false,
+                        underline: false,
+                        code: false,
+                        color: "default",
+                    },
+                    plain_text: " the end.",
+                    href: null,
+                },
+            ],
+            color: "default",
+        },
+    }, targetPage);
+    expect(results.trim()).toBe(`Inline [great page](/${targetPageId}) the end.`);
+}));
 test("link to a heading block on a page", () => __awaiter(void 0, void 0, void 0, function* () {
     const targetPageId = "123";
     const blocks = {

package/package.json

@@ -98,5 +98,5 @@
     "esbuild@0.17.17": true,
     "fsevents": false
   },
-  "version": "1.0.0-alpha.5"
+  "version": "1.0.0-alpha.7"
 }