diff options
| author | bob <bcz@cs.brown.edu> | 2020-02-12 14:54:20 -0500 |
|---|---|---|
| committer | bob <bcz@cs.brown.edu> | 2020-02-12 14:54:20 -0500 |
| commit | 2d6e00f48f7c2e6b5a8f512a6427ba56b1b216b4 (patch) | |
| tree | 0812859d6c2c6cb897f148588eb4bb2a198be11b /src/scraping | |
| parent | 30f42e403bd8cca4d20fc1b0d1cf28fbfa3a1524 (diff) | |
| parent | 2d76d891cb6a656f3c404d17ffa1cac05e08d115 (diff) | |
Merge branch 'master' of https://github.com/browngraphicslab/Dash-Web
Diffstat (limited to 'src/scraping')
| -rw-r--r-- | src/scraping/buxton/final/BuxtonImporter.ts | 9 |
1 files changed, 4 insertions, 5 deletions
diff --git a/src/scraping/buxton/final/BuxtonImporter.ts b/src/scraping/buxton/final/BuxtonImporter.ts index d9d48d68c..f9726872c 100644 --- a/src/scraping/buxton/final/BuxtonImporter.ts +++ b/src/scraping/buxton/final/BuxtonImporter.ts @@ -115,7 +115,7 @@ const RegexMap = new Map<keyof DeviceDocument, Processor<any>>([ }], ["year", { exp: /Year:\s+([^\|]*)\s+\|/, - transformer: Utilities.numberValue + transformer: (raw: string) => Utilities.numberValue(/[0-9]{4}/.exec(raw)![0]) }], ["primaryKey", { exp: /Primary:\s+(.*)(Secondary|Additional):/, @@ -254,10 +254,9 @@ async function extractFileContents(pathToDocument: string): Promise<DocumentCont strictEqual(length % 3 === 0, true, "Improper caption formatting."); for (let i = 3; i < captionTargets.length; i += 3) { - const [image, fileName, caption] = captionTargets.slice(i, i + 3); - strictEqual(image, "", `The image cell in one row was not the empty string: ${image}`); - captions.push(caption); - embeddedFileNames.push(fileName); + const row = captionTargets.slice(i, i + 3); + captions.push(row[1]); + embeddedFileNames.push(row[2]); } // extract all hyperlinks embedded in the document |
