anyfreebook-cli 0.1.0 → 0.2.0

package/README.md

@@ -148,7 +148,7 @@ anyfreebook search "python" --json | jq '.results[] | select(.tier=="download")
 | --- | --- | --- |
 | `--limit <n>` | 10 | Results per page (max 50). |
 | `--page <n>` | 1 | Page number. |
-| `--source <name>` | `all` | `all`, `gutenberg`, `openlibrary`, `archive`, `googlebooks`, `pubmed`, `doaj`, `librivox`. |
+| `--source <name>` | `all` | `all`, `gutenberg`, `openlibrary`, `archive`, `googlebooks`, `pubmed`, `doaj`, `librivox`, `openalex`, `wikisource`, `europeana`, `zenodo`, `otl`. |
 | `--format <fmt>` | — | Keep only results offering `epub`, `pdf`, `txt` or `html`. |
 | `--free-only` | off | Keep only `tier === "download"` results. |
 | `--json` | off | Print the raw JSON response to stdout and nothing else. |
@@ -157,6 +157,7 @@ anyfreebook search "python" --json | jq '.results[] | select(.tier=="download")
 
 | Option | Default | Meaning |
 | --- | --- | --- |
+| `--author <name>` | — | Resolve this exact title+author instead of searching. Recommended. |
 | `--limit <n>` | 5 | Results to consider. |
 | `--source <name>` | `all` | Same values as `search`. |
 | `--format <fmt>` | best available | `epub` > `pdf` > `txt` > `html`. |
@@ -168,6 +169,19 @@ Downloads are **always** restricted to `tier === "download"`. A `borrow` or
 `unverified` result is reported as refused, with the reason — it is never
 skipped quietly.
 
+A result is only downloaded if its **own title** is the title you asked for.
+The search fans out to sources that match free text, so "Pride and Prejudice"
+also finds real academic papers about prejudice, and one of those can outrank
+the novel and carry a perfectly good PDF. Those are reported and skipped, not
+downloaded as if they were the book.
+
+Pass `--author` when you know it. That resolves one exact title against the
+real record instead of searching, which is both faster and unambiguous:
+
+```bash
+anyfreebook download "Pride and Prejudice" --author "Jane Austen" --format epub
+```
+
 File names are `<title-slug>-<source-id>.<ext>`, so two different books with
 the same title do not overwrite each other.
 

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "anyfreebook-cli",
-  "version": "0.1.0",
-  "description": "Search 7 open-access book sources and download only what is verifiably public domain.",
+  "version": "0.2.0",
+  "description": "Search 12 open-access book sources and download only what is verifiably public domain.",
   "type": "module",
   "bin": {
     "anyfreebook": "bin/anyfreebook.js"

package/src/cli.js

@@ -92,7 +92,7 @@ export function parseArgs(argv) {
   return opts;
 }
 
-export const GENERAL_HELP = `anyfreebook ${VERSION} — search 7 open-access book sources, download only what is verifiably public domain.
+export const GENERAL_HELP = `anyfreebook ${VERSION} — search 12 open-access book sources, download only what is verifiably public domain.
 
 Usage:
   anyfreebook search <query> [options]
@@ -127,7 +127,8 @@ Archive, Google Books, PubMed Central, DOAJ and LibriVox.
   --limit <n>        Results per page (default 10, max 50)
   --page <n>         Page number (default 1)
   --source <name>    all | gutenberg | openlibrary | archive | googlebooks |
-                     pubmed | doaj | librivox   (default all)
+                     pubmed | doaj | librivox | openalex | wikisource |
+                     europeana | zenodo | otl   (default all)
   --format <fmt>     Keep only results offering epub | pdf | txt | html
   --free-only        Keep only results with a verified free download
 `,
@@ -138,7 +139,8 @@ Download — search, then write verified public-domain books to a folder.
 
   --limit <n>        Results to consider (default 5, max 50)
   --source <name>    all | gutenberg | openlibrary | archive | googlebooks |
-                     pubmed | doaj | librivox   (default all)
+                     pubmed | doaj | librivox | openalex | wikisource |
+                     europeana | zenodo | otl   (default all)
   --format <fmt>     epub | pdf | txt | html (default: best available —
                      epub > pdf > txt > html)
   --out <dir>        Destination folder (default ./books)

package/src/commands/download.js

@@ -57,6 +57,57 @@ export function refusalMessage(result) {
   return `"${title}" has tier "${result?.tier ?? 'unknown'}" — only tier=download results may be downloaded.`;
 }
 
+/** Lowercase, alphanumeric-only, single-spaced. Used to compare titles. */
+function normaliseTitle(value) {
+  return String(value || '')
+    .toLowerCase()
+    .replace(/[^a-z0-9]+/g, ' ')
+    .trim();
+}
+
+/**
+ * True when a result's own title is the title that was asked for.
+ *
+ * This matters because the search fans out to sources that match free text.
+ * "Pride and Prejudice" finds real academic papers about prejudice — one
+ * titled "assessing the societal and cultural factors influencing males
+ * pursuing nursing and midwifery" — and those papers are more numerous than
+ * the novel. Downloading the top hits without checking their titles means
+ * fetching whatever happened to rank highest.
+ *
+ * A single word is treated as a keyword and matched anywhere in the title. A
+ * multi-word phrase is treated as a title and must be the title, or the title
+ * followed by a subtitle — "Pride and Prejudice" is the novel, but "Pride and
+ * Prejudice in the air" is a different book that happens to start the same
+ * way.
+ */
+export function titleMatches(query, title) {
+  const want = normaliseTitle(query);
+  const got = normaliseTitle(title);
+  if (!want || !got) return false;
+
+  if (!want.includes(' ')) {
+    const escaped = want.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
+    return new RegExp(`\\b${escaped}\\b`).test(got);
+  }
+
+  if (got === want) return true;
+
+  const rawWant = String(query || '').toLowerCase().trim();
+  const rawGot = String(title || '').toLowerCase().trim();
+  const separator = /^[:\-–—(\[]/;
+
+  // A subtitle or an edition marker may follow the title.
+  if (rawGot.startsWith(rawWant) && separator.test(rawGot.slice(rawWant.length).trimStart())) {
+    return true;
+  }
+  // The request may be more specific than the record's title.
+  if (rawWant.startsWith(rawGot) && separator.test(rawWant.slice(rawGot.length).trimStart())) {
+    return true;
+  }
+  return false;
+}
+
 async function pathExists(target) {
   try {
     await fs.access(target);
@@ -160,13 +211,33 @@ export async function downloadCommand(opts, ctx) {
   const dryRun = Boolean(opts.dryRun);
   const outDir = path.resolve(opts.out || DEFAULT_OUT_DIR);
 
-  const data = await ctx.api.search({ q: query, page: 1, source, limit });
-  const failedSources = Array.isArray(data.failedSources) ? data.failedSources : [];
-  if (data.partial || failedSources.length > 0) {
-    ctx.err.write(`${partialWarning(failedSources, ctx.style)}\n`);
+  const author = typeof opts.author === 'string' ? opts.author.trim() : '';
+
+  // With an author, the resolve endpoint is authoritative: it matches the
+  // title AND the author against the real record, so a search that returns
+  // papers which merely contain the word "prejudice" cannot displace the
+  // novel. Without one, the search hits are filtered to the titles that
+  // actually match, for the same reason.
+  let results = [];
+  let resolvedDirectly = false;
+
+  if (author) {
+    const resolved = await ctx.api.resolve({ title: query, author });
+    if (resolved && resolved.tier === 'download') {
+      results = [resolved];
+      resolvedDirectly = true;
+    }
+  }
+
+  if (!resolvedDirectly) {
+    const data = await ctx.api.search({ q: query, page: 1, source, limit });
+    const failedSources = Array.isArray(data.failedSources) ? data.failedSources : [];
+    if (data.partial || failedSources.length > 0) {
+      ctx.err.write(`${partialWarning(failedSources, ctx.style)}\n`);
+    }
+    results = Array.isArray(data.results) ? data.results : [];
   }
 
-  const results = Array.isArray(data.results) ? data.results : [];
   const items = [];
 
   for (const result of results) {
@@ -187,6 +258,19 @@ export async function downloadCommand(opts, ctx) {
       });
       continue;
     }
+    // Only download a file when the result is the title that was asked for.
+    // The search fans out to sources that match free text, so "Pride and
+    // Prejudice" also finds real papers about prejudice — and one of those
+    // can outrank the novel and carry a perfectly downloadable PDF.
+    if (!resolvedDirectly && !titleMatches(query, result?.title)) {
+      items.push({
+        status: 'skipped',
+        title,
+        book: result,
+        message: `Not titled "${query}" — not downloading it as if it were the book you asked for.`,
+      });
+      continue;
+    }
     items.push({ status: 'pending', title, book: result, file });
   }
 

package/src/config.js

@@ -42,6 +42,11 @@ export const SOURCES = [
   'pubmed',
   'doaj',
   'librivox',
+  'openalex',
+  'wikisource',
+  'europeana',
+  'zenodo',
+  'otl',
 ];
 
 /** File formats the CLI knows how to download. */