Add spl-photo and partner-photo

internal/db/gdocs.go

@@ -77,4 +77,5 @@ type EmbedImage struct {
 	Description string `json:"description"`
 	Width       int    `json:"width"`
 	Height      int    `json:"height"`
+	Kind        string `json:"kind"`
 }

pkg/almanack/service-gdocs-markdown.go

@@ -66,6 +66,10 @@ func intermediateDocToMarkdown(doc *html.Node) string {
 		// Write picture shortcode
 		case db.ImageEmbedTag:
 			image := dbembed.Value.(db.EmbedImage)
+			if image.Kind == "partner" {
+				dataEl.Parent.RemoveChild(dataEl)
+				continue
+			}
 			var widthHeight string
 			if image.Width != 0 {
 				widthHeight = fmt.Sprintf(`width-ratio="%d" height-ratio="%d" `,

pkg/almanack/service-gdocs-process.go

@@ -134,14 +134,25 @@ func processDocHTML(docHTML *html.Node) (
 			data := newDataTag(dtPartnerText, xhtml.InnerHTMLBlocks(n))
 			xhtml.ReplaceWith(tbl, data)
 
-		case "photo", "image", "photograph", "illustration", "illo":
+		case "photo", "image", "photograph", "illustration", "illo", "spl-photo", "partner-photo":
 			embed.Type = db.ImageEmbedTag
-			if imageEmbed, warning := processImage(rows, n); warning != "" {
+			kind := "all"
+			if label == "spl-photo" {
+				kind = "spl"
+			} else if label == "partner-photo" {
+				kind = "partner"
+			}
+			if imageEmbed, warning := processImage(rows, n, kind); warning != "" {
 				tbl.Parent.RemoveChild(tbl)
 				warnings = append(warnings, warning)
 			} else {
 				embed.Value = *imageEmbed
-				goto append
+				if kind != "spl" {
+					embeds = append(embeds, embed)
+					n++
+				}
+				data := newDataTag(dtDBEmbed, dbEmbedToString(embed))
+				xhtml.ReplaceWith(tbl, data)
 			}
 
 		case "metadata", "info":
@@ -222,7 +233,7 @@ func processDocHTML(docHTML *html.Node) (
 	return
 }
 
-func processImage(rows xhtml.TableNodes, n int) (imageEmbed *db.EmbedImage, warning string) {
+func processImage(rows xhtml.TableNodes, n int, kind string) (imageEmbed *db.EmbedImage, warning string) {
 	var width, height int
 	if w := xhtml.TextContent(rows.Value("width")); w != "" {
 		width, _ = strconv.Atoi(w)
@@ -239,6 +250,7 @@ func processImage(rows xhtml.TableNodes, n int) (imageEmbed *db.EmbedImage, warn
 		),
 		Width:  width,
 		Height: height,
+		Kind:   kind,
 	}
 
 	if path := xhtml.TextContent(rows.Value("path")); path != "" {

pkg/almanack/service-gdocs-raw.go

@@ -30,6 +30,10 @@ func intermediateDocToPartnerHTML(doc *html.Node) {
 		switch dbembed.Type {
 		// Replace images with red placeholder text
 		case db.ImageEmbedTag:
+			if imgTag := dbembed.Value.(db.EmbedImage); imgTag.Kind == "spl" {
+				dataEl.Parent.RemoveChild(dataEl)
+				continue
+			}
 			placeholder := xhtml.New("h2", "style", "color: red;")
 			xhtml.AppendText(placeholder, fmt.Sprintf("Embed #%d", dbembed.N))
 			xhtml.ReplaceWith(dataEl, placeholder)

pkg/almanack/service-gdocs-rich.go

@@ -3,6 +3,7 @@ package almanack
 import (
 	"fmt"
 
+	"github.com/spotlightpa/almanack/internal/db"
 	"github.com/spotlightpa/almanack/internal/iterx"
 	"github.com/spotlightpa/almanack/internal/xhtml"
 	"golang.org/x/net/html"
@@ -27,6 +28,10 @@ func intermediateDocToPartnerRichText(doc *html.Node) {
 	// Replace other embeds with red placeholder text
 	for dataEl, value := range dataEls(doc, dtDBEmbed) {
 		dbembed := dbEmbedFromString(value)
+		if imgTag, ok := dbembed.Value.(db.EmbedImage); ok && imgTag.Kind == "spl" {
+			dataEl.Parent.RemoveChild(dataEl)
+			continue
+		}
 		placeholder := xhtml.New("h2", "style", "color: red;")
 		xhtml.AppendText(placeholder, fmt.Sprintf("Embed #%d", dbembed.N))
 		xhtml.ReplaceWith(dataEl, placeholder)

pkg/almanack/service-gdocs.go

@@ -113,7 +113,7 @@ func (svc Services) ProcessGDocsDoc(ctx context.Context, dbDoc db.GDocsDoc) (err
 	// Handle image uploads/database lookups
 	for tbl, rows := range xhtml.Tables(docHTML) {
 		switch label := rows.Label(); label {
-		case "photo", "image", "photograph", "illustration", "illo":
+		case "photo", "image", "photograph", "illustration", "illo", "spl-photo", "partner-photo":
 			if warning := svc.replaceImagePath(
 				ctx, tbl, rows, dbDoc.ExternalID, objID2Path,
 			); warning != "" {

pkg/almanack/testdata/processDocHTML/OP1/embeds.json

@@ -13,7 +13,8 @@
       "caption": "Here is a caption",
       "description": "Blah blah blah.",
       "width": 0,
-      "height": 0
+      "height": 0,
+      "kind": "all"
     }
   },
   {
@@ -25,7 +26,8 @@
       "caption": "Overdose awareness",
       "description": "",
       "width": 0,
-      "height": 0
+      "height": 0,
+      "kind": "all"
     }
   },
   {
@@ -37,7 +39,8 @@
       "caption": "",
       "description": "The House floor in the Pa. Capitol. The chamber is controlled by Democrats for the first session in more than a \"decade.\"",
       "width": 0,
-      "height": 0
+      "height": 0,
+      "kind": "all"
     }
   }
-]
+]

pkg/almanack/testdata/processDocHTML/OP2/article.md

@@ -0,0 +1,19 @@
+My name is \*<a href="mailto:[email protected]">Carlana Johnson</a>\*.
+
+This is my \_<a href="https://docs.google.com/document/d/103kCeBG2OQS_ZHkHUyKpT9Z_ajs4tuQ-WtCvlj79Vqs/edit">test document</a>\_.
+
+\[Citation Needed\]
+
+<strong>Lorem ipsum</strong> dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim <em>id est laborum</em>.
+
+# Some blocks
+
+{{<embed/raw srcdoc="&lt;script src=&#34;http://example.com/&#34;&gt;&lt;/script&gt;">}}
+
+Here is <em>some</em> <strong>text</strong>. Lorem ipsum.
+
+{{<picture src="external/75h9dj4qkszh2ydb6e3h9pn2r0.jpeg" description="" caption="Overdose awareness" credit="Amanda Berg">}}
+
+And here’s another image:
+
+{{<picture src="external/08v6tmkadg07fyknwy3pb8sm3m.jpeg" description="The House floor in the Pa. Capitol. The chamber is controlled by Democrats for the first session in more than a &#34;decade.&#34;" caption="" credit="Amanda Berg / For Spotlight PA">}}

pkg/almanack/testdata/processDocHTML/OP2/doc.html

@@ -0,0 +1,86 @@
+<p>
+</p><table><tr><td><p>metadata
+</p></td><td><p>
+</p></td></tr><tr><td><p><strong>Slug
+</strong></p></td><td><p>OP1
+</p></td></tr><tr><td><p><strong>Date
+</strong></p></td><td><p>4/20/2023
+</p></td></tr><tr><td><p><strong>Hed
+</strong></p></td><td><p>The Winds of Change
+</p></td></tr><tr><td><p><strong>Description
+</strong></p></td><td><p>Something is in the air tonight…
+</p></td></tr><tr><td><p><strong>Byline
+</strong></p></td><td><p>The Scorpions
+</p></td></tr><tr><td><p><strong>Blurb (100 chars)
+</strong></p></td><td><p>Blurb
+</p></td></tr><tr><td><p><strong>Budget
+</strong></p></td><td><p>
+</p></td></tr><tr><td><p>A CIA op becomes an unexpected smash hit.<strong>
+</strong></p></td><td><p>
+</p></td></tr><tr><td><p><strong>Lede image
+</strong></p></td><td><p>N / A
+</p></td></tr><tr><td><p><strong>Lede image credit
+</strong></p></td><td><p>Teenage Engineering
+</p></td></tr><tr><td><p><strong>Lede image description
+</strong></p></td><td><p>Synthesizer
+</p></td></tr></table><p>
+</p><p>My name is *<a href="mailto:[email protected]">Carlana Johnson</a>*.
+</p><p>
+</p><p>This is my _<a href="https://docs.google.com/document/d/103kCeBG2OQS_ZHkHUyKpT9Z_ajs4tuQ-WtCvlj79Vqs/edit">test document</a>_.
+</p><p>
+</p><p>[Citation Needed]
+</p><p>
+</p><p><strong>Lorem ipsum</strong> dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim <em>id est laborum</em>.
+</p><p>
+</p><h1>Some blocks
+</h1><p>
+</p><table><tr><td><p>html
+</p></td></tr><tr><td><p>
+</p><p>&lt;script src=&#34;http://example.com/&#34;&gt;&lt;/script&gt;
+</p><p>
+</p></td></tr></table><p>
+</p><p>
+</p><table><tr><td><p>partner-photo
+</p></td><td><p>
+</p></td></tr><tr><td><p><a href="https://carlmjohnson.net/images/dither-320.png"><img src="https://lh6.googleusercontent.com/bOd9ROs85eAzE7mvH2ySxfKVSC98pIdaGOZiaOMFZRhNalaxVzOv7zdlWFV4AKNe-YhurZgLL4Jxe_bNJ_xFblXcvMg3Fs0P-B7azYu5iPhOlFEg2YrJCp83QhsLhe5ahrUMNDoBWre4clZ8NtrVu6Jq1wMPIVT2" title="" alt="" data-oid="kix.4v7yhj6ax6xk"/></a>
+</p></td><td><p>
+</p></td></tr><tr><td><p>credit
+</p></td><td><p>
+</p></td></tr><tr><td><p>CarlCo
+</p></td><td><p>
+</p></td></tr><tr><td><p>description
+</p></td><td><p>
+</p></td></tr><tr><td><p>Blah blah blah.
+</p></td><td><p>
+</p></td></tr><tr><td><p>caption
+</p></td><td><p>
+</p></td></tr><tr><td><p>Here is a caption
+</p></td><td><p>
+</p></td></tr><tr><td>path</td><td>external/cwp3y3z5x8svbtsfv7t68hggc0.png</td></tr></table><p>
+</p><p>Here is <em>some</em> <strong>text</strong>. Lorem ipsum.
+</p><p>
+</p><p>
+</p><table><tr><td><p><strong>spl-photo
+</strong></p></td></tr><tr><td><p><a href="https://drive.google.com/file/d/1ssiQd8AKXHo99qkZZwYbHxfVJHY3RPnL/view?usp=share_link">20210831_spotlight_overdose_awareness_selects_0021.JPG</a>
+</p></td></tr><tr><td><p>credit
+</p></td></tr><tr><td><p>Amanda Berg
+</p></td></tr><tr><td><p>caption
+</p></td></tr><tr><td><p>Overdose awareness
+</p></td></tr><tr><td>path</td><td>external/75h9dj4qkszh2ydb6e3h9pn2r0.jpeg</td></tr></table><p>
+</p><p>And here’s another image:
+</p><p>
+</p><table><tr><td><p><strong>photo
+</strong></p></td><td><p>
+</p></td></tr><tr><td><p>path
+</p></td><td><p><mark>external/08v6tmkadg07fyknwy3pb8sm3m.jpeg</mark>
+</p></td></tr><tr><td><p>credit
+</p></td><td><p>
+</p></td></tr><tr><td><p>Amanda Berg / For Spotlight PA
+</p></td><td><p>
+</p></td></tr><tr><td><p>Alt
+</p></td><td><p>
+</p></td></tr><tr><td><p>The House floor in the Pa. Capitol. The chamber is controlled by Democrats for the first session in more than a &#34;decade.&#34;
+</p></td><td><p>
+</p></td></tr></table><p>
+</p><p>
+</p>

pkg/almanack/testdata/processDocHTML/OP2/embeds.json

@@ -0,0 +1,33 @@
+[
+  {
+    "n": 1,
+    "type": "raw",
+    "value": "<script src=\"http://example.com/\"></script>"
+  },
+  {
+    "n": 2,
+    "type": "image",
+    "value": {
+      "path": "external/cwp3y3z5x8svbtsfv7t68hggc0.png",
+      "credit": "CarlCo",
+      "caption": "Here is a caption",
+      "description": "Blah blah blah.",
+      "width": 0,
+      "height": 0,
+      "kind": "partner"
+    }
+  },
+  {
+    "n": 3,
+    "type": "image",
+    "value": {
+      "path": "external/08v6tmkadg07fyknwy3pb8sm3m.jpeg",
+      "credit": "Amanda Berg / For Spotlight PA",
+      "caption": "",
+      "description": "The House floor in the Pa. Capitol. The chamber is controlled by Democrats for the first session in more than a \"decade.\"",
+      "width": 0,
+      "height": 0,
+      "kind": "all"
+    }
+  }
+]

pkg/almanack/testdata/processDocHTML/OP2/metadata.json

@@ -0,0 +1,19 @@
+{
+  "publication_date": null,
+  "internal_id": "OP1",
+  "byline": "The Scorpions",
+  "budget": "A CIA op becomes an unexpected smash hit.",
+  "hed": "The Winds of Change",
+  "description": "Something is in the air tonight…",
+  "lede_image": "",
+  "lede_image_credit": "Teenage Engineering",
+  "lede_image_description": "Synthesizer",
+  "lede_image_caption": "",
+  "eyebrow": "",
+  "url_slug": "",
+  "blurb": "Blurb",
+  "link_title": "",
+  "seo_title": "",
+  "og_title": "",
+  "twitter_title": ""
+}

pkg/almanack/testdata/processDocHTML/OP2/raw.html

@@ -0,0 +1 @@
+<body><p>My name is *<a href="mailto:[email protected]">Carlana Johnson</a>*.</p><p>This is my _<a href="https://docs.google.com/document/d/103kCeBG2OQS_ZHkHUyKpT9Z_ajs4tuQ-WtCvlj79Vqs/edit">test document</a>_.</p><p>[Citation Needed]</p><p><strong>Lorem ipsum</strong> dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim <em>id est laborum</em>.</p><h1>Some blocks</h1><script src="http://example.com/"></script><h2 style="color: red;">Embed #2</h2><p>Here is <em>some</em> <strong>text</strong>. Lorem ipsum.</p><p>And here’s another image:</p><h2 style="color: red;">Embed #3</h2></body>

pkg/almanack/testdata/processDocHTML/OP2/rich.html

@@ -0,0 +1 @@
+<body><p>My name is *<a href="mailto:[email protected]">Carlana Johnson</a>*.</p><p>This is my _<a href="https://docs.google.com/document/d/103kCeBG2OQS_ZHkHUyKpT9Z_ajs4tuQ-WtCvlj79Vqs/edit">test document</a>_.</p><p>[Citation Needed]</p><p><strong>Lorem ipsum</strong> dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim <em>id est laborum</em>.</p><h1>Some blocks</h1><h2 style="color: red;">Embed #1</h2><h2 style="color: red;">Embed #2</h2><p>Here is <em>some</em> <strong>text</strong>. Lorem ipsum.</p><p>And here’s another image:</p><h2 style="color: red;">Embed #3</h2></body>

pkg/almanack/testdata/processDocHTML/OP2/warnings.json

@@ -0,0 +1 @@
+null

pkg/almanack/testdata/processDocHTML/SPLHAROLD/embeds.json

@@ -13,7 +13,8 @@
       "caption": "A plaque inside Millvale \"they bar\" Harold's Haunt profiles the businesses' owners and namesake, a neighborhood ghost.",
       "description": "A plaque inside Harold’s Haunt is pictured.",
       "width": 0,
-      "height": 0
+      "height": 0,
+      "kind": "all"
     }
   }
-]
+]

pkg/almanack/testdata/processDocHTML/abc/embeds.json

@@ -8,7 +8,8 @@
       "caption": "Cap 1",
       "description": "Desc 2",
       "width": 640,
-      "height": 480
+      "height": 480,
+      "kind": "all"
     }
   }
-]
+]