Add publisher-meta scraper fallback + 3 Springer cache refreshes (#59)

Drive `just validate-references-all` errors from 43 to 39 (and from
session-start 101 to 39) by adding a last-resort DOI page scraper to
the literature fetcher and refreshing the three Springer caches it
unblocks.

Fetcher (src/communitymech/literature.py):
- fetch_publisher_meta_abstract(): GET https://doi.org/<DOI>, follow
  redirects, and pull the abstract excerpt out of the page's
  twitter:description / og:description / description meta tag. Springer
  publishes the first ~200 characters of the abstract in
  twitter:description even for paywalled articles where Crossref /
  OpenAlex / Semantic Scholar / Europe PMC have no abstract. Includes
  on-disk caching as publisher_<safe-doi>.txt and strips the
  "Journal Name - " prefix Springer adds to that field. Elsevier
  ScienceDirect intentionally serves a bot-detection page and yields
  nothing - that's the residual cap.
- fetch_paper() fallback chain now: CrossRef -> PMID -> PMC -> OpenAlex
  -> Semantic Scholar -> Europe PMC -> publisher meta scrape.

Cache refresh (recovers 4 ERROR rows):
- DOI_10.1007_s10311-019-00911-y (Ewaste copper bioleaching, Springer)
- DOI_10.1007_s10230-008-0059-z (Iberian meromictic pit lakes, Springer)
- DOI_10.1007_BF02106205 (Acidobacterium taxonomy paper, Current
  Microbiology / Springer; cited 2x in AMD_Acidophile_Heterotroph_Network)

Snippet repairs:
- Ewaste_Bioleaching_Consortium: replace title quote with the abstract's
  verbatim e-waste bioleaching framing.
- Iberian_Pit_Lake_Stratified_Community: upgrade PARTIAL to SUPPORT and
  expand the snippet to the abstract's vertical-gradient quote.
- AMD_Acidophile_Heterotroph_Network: replace two title quotes with the
  abstract's verbatim genus proposal.

Remaining 39 "No content available" errors are all Elsevier 2024-2025
papers (j.jece.2025.120403, j.cej.2024.153492, j.ibiod.2025.106190,
10889868.2024.2407240) plus one ResearchGate preprint - their abstracts
are not in any aggregator we can query and the publisher pages serve
bot-detection HTML.

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: realmarcin

kb/communities/AMD_Acidophile_Heterotroph_Network.yaml

@@ -229,8 +229,8 @@ taxonomy:
   - reference: doi:10.1007/BF02106205
     supports: SUPPORT
     evidence_source: IN_VITRO
-    snippet: 'Acidobacterium capsulatum gen. nov., sp. nov.: An acidophilic chemoorganotrophic bacterium
-      containing menaquinone from acidic mineral environment'
+    snippet: Acidobacterium is proposed as a new genus for the acidophilic, chemoorganotrophic
+      bacteria containing menaquinone isolated from acidic mineral environments
     explanation: Establishes isolation from acidic mineral environment
   - reference: doi:10.3389/fmicb.2016.00744
     supports: SUPPORT
@@ -622,8 +622,8 @@ ecological_interactions:
   - reference: doi:10.1007/BF02106205
     supports: SUPPORT
     evidence_source: IN_VITRO
-    snippet: 'Acidobacterium capsulatum gen. nov., sp. nov.: An acidophilic chemoorganotrophic bacterium
-      containing menaquinone from acidic mineral environment'
+    snippet: Acidobacterium is proposed as a new genus for the acidophilic, chemoorganotrophic
+      bacteria containing menaquinone isolated from acidic mineral environments
     explanation: Establishes chemoorganotrophic metabolism
   - reference: doi:10.3389/fmicb.2016.00744
     supports: SUPPORT

kb/communities/Ewaste_Bioleaching_Consortium.yaml

@@ -136,9 +136,11 @@ taxonomy:
   - reference: doi:10.1007/s10311-019-00911-y
     supports: SUPPORT
     evidence_source: IN_VITRO
-    snippet: Enhanced bioleaching of copper from circuit boards of computer waste by Acidithiobacillus
-      ferrooxidans
-    explanation: Establishes At. ferrooxidans efficacy for e-waste copper recovery
+    snippet: Computer circuit boards are a major electronic waste containing higher concentrations
+      of copper, gold and silver. These metals may be recovered by bioleaching
+    explanation: Establishes bioleaching of computer circuit boards for recovery of copper
+      and other metals; the paper specifically evaluates Acidithiobacillus ferrooxidans
+      efficacy.
 - taxon_term:
     preferred_term: Acidithiobacillus thiooxidans
     term:

kb/communities/Iberian_Pit_Lake_Stratified_Community.yaml

@@ -697,10 +697,12 @@ environmental_factors:
       with Proteobacteria
     explanation: Documents meromictic stability
   - reference: doi:10.1007/s10230-008-0059-z
-    supports: PARTIAL
+    supports: SUPPORT
     evidence_source: IN_VIVO
-    snippet: permanently stratified acidic pit lake
-    explanation: Documents thermal stratification maintaining meromixis
+    snippet: A marked vertical trend of increasing temperature and dissolved metal concentrations
+      is observed in the monimolimnia of some meromictic pit lakes of the Iberian
+    explanation: Documents the vertical temperature and metal gradients that maintain
+      meromictic stratification in Iberian acidic pit lakes.
 - name: Microbial Diversity Gradient
   value: 72-206
   unit: ASVs per sample

references_cache/DOI_10.1007_BF02106205.md

@@ -8,7 +8,7 @@ authors:
 journal: Current Microbiology
 year: '1991'
 doi: 10.1007/BF02106205
-content_type: unavailable
+content_type: abstract_only
 ---
 
 # Acidobacterium capsulatum gen. nov., sp. nov.: An acidophilic chemoorganotrophic bacterium containing menaquinone from acidic mineral environment
@@ -17,3 +17,5 @@ content_type: unavailable
 **DOI:** [10.1007/BF02106205](https://doi.org/10.1007/BF02106205)
 
 ## Content
+
+Acidobacterium is proposed as a new genus for the acidophilic, chemoorganotrophic bacteria containing menaquinone isolated from acidic mineral environments.Acidobacterium

references_cache/DOI_10.1007_s10230-008-0059-z.md

@@ -9,7 +9,7 @@ authors:
 journal: Mine Water and the Environment
 year: '2009'
 doi: 10.1007/s10230-008-0059-z
-content_type: unavailable
+content_type: abstract_only
 ---
 
 # Physico-chemical gradients and meromictic stratification in Cueva de la Mora and other acidic pit lakes of the Iberian Pyrite Belt
@@ -18,3 +18,5 @@ content_type: unavailable
 **DOI:** [10.1007/s10230-008-0059-z](https://doi.org/10.1007/s10230-008-0059-z)
 
 ## Content
+
+A marked vertical trend of increasing temperature and dissolved metal concentrations is observed in the monimolimnia of some meromictic pit lakes of the Iberian

references_cache/DOI_10.1007_s10311-019-00911-y.md

@@ -7,7 +7,7 @@ authors:
 journal: Environmental Chemistry Letters
 year: '2019'
 doi: 10.1007/s10311-019-00911-y
-content_type: unavailable
+content_type: abstract_only
 ---
 
 # Enhanced bioleaching of copper from circuit boards of computer waste by Acidithiobacillus ferrooxidans
@@ -16,3 +16,5 @@ content_type: unavailable
 **DOI:** [10.1007/s10311-019-00911-y](https://doi.org/10.1007/s10311-019-00911-y)
 
 ## Content
+
+Computer circuit boards are a major electronic waste containing higher concentrations of copper, gold and silver. These metals may be recovered by bioleaching, an

src/communitymech/literature.py

@@ -368,6 +368,72 @@ def fetch_europepmc_abstract(self, doi: str) -> str | None:
             print(f"Error fetching Europe PMC for {doi}: {e}")
             return None
 
+    def fetch_publisher_meta_abstract(self, doi: str) -> str | None:
+        """
+        Last-resort scrape of the DOI landing page for an abstract excerpt.
+
+        Most publishers expose the abstract (or its first ~200 chars) in
+        page-level meta tags - typically `twitter:description`,
+        `og:description`, or the standard `<meta name="description">` -
+        even when the article itself is paywalled and Crossref / OpenAlex /
+        Semantic Scholar / Europe PMC carry no abstract. Springer/Nature
+        is the most reliable source for this pattern; Elsevier
+        ScienceDirect serves a bot-detection page and yields nothing.
+
+        Args:
+            doi: DOI string (with or without "doi:" prefix)
+
+        Returns:
+            Abstract excerpt (may be truncated to ~200 chars by the
+            publisher) or None.
+        """
+        doi = re.sub(r"^(?i:doi:)", "", doi).replace("https://doi.org/", "").strip()
+        cache_file = self._abstract_cache_path("publisher", doi)
+        if cache_file.exists():
+            return cache_file.read_text() or None
+
+        try:
+            response = self.session.get(
+                f"https://doi.org/{doi}",
+                headers={"User-Agent": "Mozilla/5.0 (Macintosh)"},
+                allow_redirects=True,
+                timeout=30,
+            )
+            response.raise_for_status()
+            html = response.text
+        except requests.exceptions.RequestException as e:
+            print(f"Error scraping publisher page for {doi}: {e}")
+            return None
+
+        for tag in ("twitter:description", "og:description", "description"):
+            match = re.search(
+                rf'<meta\s+[^>]*name=["\']?{tag}["\']?\s+content=["\']([^"\']+)["\']',
+                html,
+                re.IGNORECASE,
+            )
+            if not match:
+                match = re.search(
+                    rf'<meta\s+[^>]*property=["\']?{tag}["\']?\s+content=["\']([^"\']+)["\']',
+                    html,
+                    re.IGNORECASE,
+                )
+            if match:
+                text = match.group(1)
+                # Decode common HTML entities
+                text = (
+                    text.replace("&amp;", "&")
+                    .replace("&quot;", '"')
+                    .replace("&#x27;", "'")
+                    .replace("&lt;", "<")
+                    .replace("&gt;", ">")
+                )
+                # Strip Springer's "Journal Name - " prefix and trailing ellipsis
+                text = re.sub(r"^[^-]+-\s*", "", text).rstrip(" .")
+                if len(text) > 80:  # skip nav-text and similar short snippets
+                    cache_file.write_text(text)
+                    return text
+        return None
+
     def fetch_unpaywall(self, doi: str, email: str = "noreply@example.com") -> str | None:
         """
         Try to fetch open access PDF URL from Unpaywall.
@@ -458,6 +524,8 @@ def fetch_paper(
                 abstract = self.fetch_semantic_scholar_abstract(doi)
             if not abstract:
                 abstract = self.fetch_europepmc_abstract(doi)
+            if not abstract:
+                abstract = self.fetch_publisher_meta_abstract(doi)
 
             return (abstract, pdf_url)