Metadata Enrichment

RSS feeds often provide incomplete article metadata.

One publisher may supply a full abstract, DOI, authors, publication date, and subjects, while another may provide only a title and URL.

RSSrch Metadata Enrichment supplements sparse Zotero feed items with additional research-useful metadata when trustworthy information can be resolved.

Possible improvements include:

DOI recovery
abstract improvement
creator recovery
publication and publisher metadata
publication date
section or article type
volume, issue, and page information
ISSN and language
subject and keyword metadata
other structured article metadata

Metadata Enrichment is enabled by default.

Enrichment supplements Zotero feeds

Zotero remains responsible for:

RSS subscriptions
feed retrieval
native feed-item storage
feed refresh
native item management

RSSrch adds an enrichment layer on top of that data.

Conceptually:

Zotero feed item
↓
Native feed metadata
+
RSSrch enrichment overlay
↓
Effective research metadata

Most enriched metadata is stored in RSSrch’s persistent enrichment data rather than written back over Zotero’s original feed fields.

This allows RSSrch to improve research presentation, rules, and scoring without taking ownership of Zotero’s feed records.

Enrichment does not create My Library items

Metadata Enrichment operates on Zotero feed items.

It does not automatically create normal Zotero library items.

For example:

Feed item
↓
Metadata Enrichment
↓
Still a feed item

Adding a paper to My Library is a separate Zotero operation.

RSSrch can invoke that operation through Actions or through the Reader’s Add to My Library control, but enrichment itself does not perform it.

Feed metadata normally remains authoritative

RSSrch does not blindly replace feed metadata whenever an external source returns another value.

For ordinary scalar metadata fields, an existing useful native feed value generally takes precedence.

For example:

Native publication title exists
↓
Use native publication title
 
Native publication title missing
+
enrichment provides one
↓
Use enriched publication title

This applies to fields such as publication metadata and date.

There are two important quality-aware exceptions:

abstracts
creators

RSSrch evaluates those fields more carefully because some feeds technically provide them but provide poor or malformed values.

Abstract quality is evaluated

A feed-provided abstract is not treated as useful merely because the field is non-empty.

RSSrch evaluates abstract quality using characteristics such as:

length
word count
title duplication
obvious boilerplate
whether the text appears to be only a short summary fragment

Internally, abstracts are classified broadly as:

missing
weak
brief
usable
substantial

Examples of weak values can include:

Abstract not available
 
View full text
 
Copyright ...
 
[article title repeated as the abstract]

RSSrch can reject an enrichment candidate that is itself too weak or looks like boilerplate.

Better abstracts can replace weak feed abstracts

If the native feed abstract is already useful, RSSrch normally keeps it.

An enrichment candidate replaces the native abstract only when the candidate is materially better according to the quality comparison.

Conceptually:

Useful feed abstract
+
similar enrichment abstract
↓
Keep feed abstract

but:

Very short or weak feed abstract
+
substantially better article abstract
↓
Use enriched abstract

and:

Missing feed abstract
+
usable enrichment abstract
↓
Use enriched abstract

This prevents enrichment from replacing useful feed text simply because another source happens to return an abstract.

Creator recovery is also quality-aware

Author metadata can also be malformed in RSS feeds.

RSSrch evaluates creator lists for issues such as:

missing creators
placeholder-heavy creator lists
unusable creator information
apparent first-name / last-name inversion involving initials

If the native creator list is useful, RSSrch retains it.

If it is missing or clearly lower quality, an enrichment-derived creator list can be used instead.

Conceptually:

Good feed creators
↓
retain feed creators

Missing or poor feed creators
+
better enriched creators
↓
use enriched creators

RSSrch can also repair a recognizable inverted-initial pattern for effective feed presentation when appropriate.

The enrichment overlay

A successful enrichment record can retain metadata such as:

Abstract
DOI
Publication / Source
Publisher
Section / Article Type
Date
Volume
Issue
Pages
ISSN
Language
Rights
Creators
Subjects / Keywords

RSSrch also records provenance for resolved fields.

This allows the system to distinguish information obtained from sources such as:

feed
publisher resolver
structured page metadata
PubMed / Europe PMC
Zotero DOI search
Zotero web translator

The RSSrch Reader can indicate that an item contains enriched metadata and identify the resolver or resolver chain that contributed it.

Automatic subject tags are different

Most enrichment values remain in the RSSrch overlay.

Accepted subject and keyword metadata can also be applied to the Zotero feed item as automatic Zotero tags when Zotero automatic tags are enabled.

Conceptually:

Publisher / structured subject metadata
↓
RSSrch validates subject term
↓
RSSrch enrichment overlay
↓
Zotero automatic tag

Those automatic tags become available through RSSrch’s:

Feed Subjects

field.

See Feed Subjects and Tags for the distinction between:

Feed Subjects
Tags
Manual Tags

What triggers automatic enrichment

RSSrch does not automatically perform an expensive metadata lookup merely because any possible metadata field is empty.

Automatic enrichment focuses on research-useful deficiencies.

An item can be considered sparse when it has problems such as:

Abstract missing
Abstract weak
Abstract brief
 
Creators missing
Creators low quality
 
Publication / Source missing
 
Publication date missing
 
Subject-tag enrichment unresolved
Subject tags discovered but not yet applied

The item must also provide enough external identity for lookup through:

an article URL
a DOI

If neither is available, RSSrch does not attempt normal automatic external enrichment.

Some fields are recovered opportunistically

Not every missing field independently triggers an automatic enrichment request.

For example, a missing:

DOI
Publisher
Section
Volume
Issue
Pages
ISSN
Language

does not necessarily cause an otherwise complete feed item to enter the automatic sparse-item queue by itself.

However, if the item is being enriched for another qualifying reason, RSSrch can recover those additional fields opportunistically.

For example:

Weak abstract
↓
Item requires enrichment
↓
Resolver also finds:
DOI
Publisher
Volume
Issue
Subjects
↓
Those useful values can be retained too

This keeps automatic network processing focused on metadata gaps that materially affect the research workflow.

New feed items are enriched automatically when needed

When Zotero receives a new feed item, RSSrch examines it through the new-item processing path.

Conceptually:

New Zotero feed item
↓
RSSrch Active?
↓
Metadata Enrichment enabled?
↓
Does item need enrichment?
│
├── No → continue RSSrch processing
│
└── Yes
    ↓
    Queue metadata enrichment
    ↓
    Resolve available metadata
    ↓
    Refresh evaluation data
    ↓
    Admission
    ↓
    Relevance Score
    ↓
    Actions

This ordering is important.

If a new item is queued for enrichment, RSSrch waits for that enrichment operation before executing Actions.

The item is then evaluated with the best effective metadata available from that processing pass.

RSSrch Active also controls automatic enrichment

Automatic enrichment is part of RSSrch processing.

If a feed’s effective RSSrch Active state is disabled:

Zotero refreshes feed
↓
New feed item stored normally
↓
RSSrch processing disabled
↓
No automatic RSSrch enrichment

Zotero continues to manage the subscription and refresh it normally.

See Feed Folders for the RSSrch Active inheritance model.

Enrichment can affect Admission and Score

Metadata Enrichment is not merely cosmetic.

Fields recovered by enrichment can participate in RSSrch rules.

For example, enrichment might provide:

Abstract:
ferroptosis is associated with...
 
Feed Subject:
Oxidative Stress
 
Publication:
Journal of Neurochemistry

Those values can then affect:

Include and Exclude Rules
Relevance Scoring
Reader highlights
Score breakdowns
subsequent Actions

After queued enrichment finishes, RSSrch invalidates the item’s prior evaluation data so rules and scoring can use the updated metadata.

Enrichment and Actions

For newly added feed items, Actions wait when enrichment has been queued.

For example:

New item arrives
↓
Abstract is missing
↓
Enrichment queued
↓
Better abstract resolved
↓
Admission evaluated
↓
Final Score calculated
↓
Action conditions evaluated

This prevents an Action from firing prematurely based only on the sparse initial RSS payload when RSSrch already knows that additional metadata resolution is pending.

If enrichment finds no additional information, normal RSSrch processing still continues using the metadata that is available.

Open:

Tools → RSSrch → Enrichment

The normal release menu contains:

Enrich Feed Metadata + Subjects
Enrich Selected Feed Metadata Now
Enrich Feed Metadata + Subjects Now…
Backfill Feed Subject Tags…

Each command serves a different purpose.

Enrich Feed Metadata + Subjects

The checkbox:

Enrich Feed Metadata + Subjects

controls normal automatic/background metadata enrichment.

It is enabled by default.

When enabled, RSSrch can automatically process sparse new feed items and maintain enrichment data.

When disabled, new automatic enrichment work is not queued.

Disabling enrichment does not erase enrichment data that RSSrch has already resolved.

It also does not remove automatic subject tags that have already been applied.

Existing enrichment information therefore remains available even when future automatic enrichment is turned off.

Disabling enrichment cancels queued work

When automatic enrichment is disabled, RSSrch removes work that is still waiting in the enrichment queue.

An operation that is already actively resolving may still finish.

Conceptually:

Disable enrichment
↓
Queued items cancelled
↓
No new automatic work queued
 
Already active lookup
↓
may finish

This avoids unnecessarily continuing a large waiting queue after the user has disabled the feature.

Enabling enrichment can perform a bounded catch-up scan

When Metadata Enrichment is turned back on, RSSrch schedules a bounded scan for sparse feed metadata.

The scan considers RSSrch-active feeds and prioritizes eligible items requiring enrichment.

It does not launch an unbounded synchronous whole-library refresh.

This allows enrichment to resume progressively without blocking Zotero with a full rescan.

Background priority

The automatic queue prioritizes items according to research usefulness and metadata need.

Higher priority is given to factors such as:

unread items
missing abstracts
weak abstracts
brief abstracts
missing or poor creators
missing publication information
missing publication date
unresolved subject-tag enrichment

This means an unread article with no useful abstract normally receives attention before an older item with only a minor metadata deficiency.

Enrich Selected Feed Metadata Now

To explicitly refresh selected feed items, use:

Tools → RSSrch → Enrichment → Enrich Selected Feed Metadata Now

The same operation is also available from the feed-item context menu as:

Enrich Feed Metadata Now

Select one or more Zotero feed items first.

This command performs a forced enrichment refresh of the selected items.

Unlike normal automatic processing, it does not require the item to pass the sparse-metadata test.

This makes it useful when:

metadata appears incorrect or incomplete
a resolver has recently improved
you want to retry a specific article
you want the strongest available manual enrichment path

Manual enrichment can run even when automatic enrichment is disabled

The selected-item command is an explicit user request.

For that reason, it uses the forced enrichment path.

The automatic Enrich Feed Metadata + Subjects toggle controls normal background processing, but it does not prevent an explicit selected-item forced refresh.

So:

Automatic enrichment:
OFF
 
User chooses:
Enrich Selected Feed Metadata Now
 
↓
explicit enrichment can still run

This allows automatic enrichment to remain disabled while individual papers are enriched on demand.

Single-item manual enrichment has an additional fallback

RSSrch deliberately avoids expensive browser-based translation during automatic background enrichment.

However, when one feed item is explicitly selected for manual enrichment, RSSrch can use Zotero’s browser-based web translator as a final fallback if lighter methods have not resolved the important metadata gaps.

Conceptually:

Single selected item
↓
DOI / structured / publisher resolvers
↓
Important gaps still remain?
↓
Zotero web translator fallback

This is reserved for a deliberate single-item operation because browser translation is substantially more expensive than structured metadata requests.

Multi-item enrichment avoids browser translation

If multiple feed items are selected:

Multiple selected items
↓
Forced enrichment
↓
Lightweight DOI / structured / publisher resolution
↓
No HiddenBrowser web-translation fallback

This prevents a large manual selection from launching many expensive browser translation operations.

The same principle applies to normal automatic/background enrichment.

Automatic queues do not launch the browser translator.

Enrich Feed Metadata + Subjects Now

Open:

Tools → RSSrch → Enrichment → Enrich Feed Metadata + Subjects Now…

RSSrch asks you to choose a feed.

It then identifies items in that feed that:

currently need metadata enrichment
do not already have a sufficiently fresh enrichment result

Those sparse items are prioritized and added to the normal enrichment queue.

This is useful when you want to process incomplete metadata for one particular feed without forcing a refresh of every item.

Unlike Enrich Selected Feed Metadata Now, this command respects the normal sparse-item and enrichment-cache logic.

Backfill Feed Subject Tags

Open:

Tools → RSSrch → Enrichment → Backfill Feed Subject Tags…

This is a specialized enrichment maintenance operation for subject metadata.

It can:

clean rejected automatic placeholder subjects
complete stale subject-tag enrichment
retry eligible still-untagged articles through a bounded subject fallback
apply accepted terms as Zotero automatic tags

The backfill requires Metadata Enrichment to be enabled.

Manual Zotero tags are not removed or converted.

See Feed Subjects and Tags for the complete subject-tag model.

How RSSrch resolves metadata

RSSrch uses several metadata strategies.

The exact sequence can vary according to the article source and the metadata already available.

Resolution can include:

DOI metadata search
+
structured webpage metadata
+
publisher-specific resolvers
+
publisher or literature APIs
+
PubMed / Europe PMC metadata
+
manual-only Zotero web translation fallback

RSSrch combines useful results while tracking where individual fields originated.

Structured page metadata

Many scholarly article pages expose structured metadata through sources such as:

citation meta tags
Dublin Core metadata
PRISM metadata
JSON-LD
publisher keyword metadata

RSSrch contains a general structured-page resolver that can extract available research metadata without opening a full browser translator.

Possible values include:

Title
Abstract
DOI
Publication
Publisher
Date
Volume
Issue
Pages
ISSN
Language
Creators
Subjects / Keywords

This is one of the primary lightweight enrichment paths.

Publisher-specific resolvers

RSSrch also contains targeted resolver adapters for sources where a specialized path provides better or more efficient metadata.

RSSrch 1.0.0 includes specialized support involving sources such as:

Europe PMC / PubMed
bioRxiv
medRxiv
MDPI
Cell Press

These adapters can use publisher APIs, XML, PubMed records, or other structured sources instead of relying entirely on large article webpages.

For example:

bioRxiv / medRxiv
→ publisher API
 
MDPI
→ article XML
 
Europe PMC
→ PubMed metadata path

The resolver architecture is prioritized so a strong specialized source can satisfy the metadata need without unnecessary additional requests.

DOI lookup

When a DOI is available, RSSrch can use Zotero’s DOI metadata translation machinery to retrieve structured metadata.

The lookup is performed without saving an item to My Library and without downloading attachments.

Conceptually:

DOI
↓
Zotero metadata search
↓
translated metadata
↓
RSSrch enrichment overlay

The DOI path is one enrichment source among several.

On specialized publisher domains, RSSrch can defer the DOI request when a direct resolver is expected to provide the same or better metadata more efficiently.

If the specialized resolver leaves important gaps, DOI lookup can still be attempted afterward.

Identity protection

External metadata is not accepted merely because a resolver returned an article.

RSSrch performs identity checks before incorporating translated metadata.

The checks can use information such as:

DOI agreement
normalized article title
substantial title overlap

For example:

Feed DOI:
10.xxxx/example1
 
Returned DOI:
10.xxxx/example2
↓
Reject metadata

RSSrch also checks title similarity when appropriate.

This reduces the risk that a search, redirect, or publisher page produces metadata belonging to a different paper.

Metadata from multiple resolvers can be combined

One resolver does not necessarily need to provide every field.

RSSrch can combine compatible metadata from multiple sources.

For example:

Publisher resolver
→ Abstract
→ Subjects
 
DOI search
→ Publication
→ Volume
→ Issue
 
Structured page
→ Publisher
→ Language

These contributions can be merged into one enrichment record.

RSSrch records the contributing resolver or field source so provenance is not lost.

Better enrichment candidates can replace weaker enrichment candidates

When several enrichment sources provide the same field, RSSrch does not simply allow the last resolver to overwrite everything.

Metadata merging considers source quality and field-specific rules.

For example, date selection considers:

resolver priority
available date precision

Abstract selection compares text quality.

Creator selection compares creator-list quality.

This allows specialized or authoritative sources to improve earlier generic results without indiscriminately replacing useful information.

Subject terms are filtered before use

Publisher metadata can contain values that are technically classified as keywords or subjects but are not useful research tags.

RSSrch filters subject candidates before storing or applying them.

Examples of rejected material can include:

placeholder values
boilerplate
excessively sentence-like values
known source-specific noise

Accepted subjects are deduplicated case-insensitively before they are retained.

See Feed Subjects and Tags for more details.

Caching avoids unnecessary repeated requests

RSSrch maintains a persistent enrichment cache.

Before performing a normal enrichment lookup, RSSrch checks whether a suitable recent result already exists.

Possible cached states include:

success
no data
error

Successful enrichment, unsuccessful lookups, and failures use different retry behavior.

This prevents RSSrch from repeatedly requesting the same unavailable metadata every time an item is displayed or evaluated.

Manual selected-item enrichment bypasses normal freshness

The explicit:

Enrich Selected Feed Metadata Now

command is different.

It uses a forced refresh.

So:

Fresh cached result
+
normal background processing
↓
reuse cache

while:

Fresh cached result
+
explicit selected-item enrichment
↓
perform forced refresh

This gives the user a direct way to retry an article without deleting enrichment state manually.

Session request caching

During an active enrichment run, RSSrch also reuses identical requests where possible.

For example, repeated attempts to resolve the same:

URL
DOI
DOI-to-PMID mapping

can reuse session state rather than immediately repeating the same network request.

RSSrch also remembers certain failed URLs and domains during the active session to reduce repeated requests to sources that are currently unavailable or blocking access.

Network requests are paced

Enrichment uses bounded request lanes rather than launching all external lookups simultaneously.

Separate request classes can be paced independently for areas such as:

DOI lookup
structured metadata
publisher APIs
NCBI / PubMed APIs
browser translation

The enrichment queue itself is also concurrency-limited.

This is intended to keep background metadata work from overwhelming Zotero or external services.

Progress reporting

When queued or manual enrichment is running, RSSrch reports progress.

The status can distinguish outcomes such as:

enriched
cached
no-data
errors
skipped

For a long operation, the progress display also reports how many items have completed and how many remain queued or active.

This is particularly useful for:

feed-level enrichment
selected multi-item enrichment
subject-tag backfill

No-data does not necessarily mean failure

A resolver can complete successfully without finding information that improves the item.

RSSrch distinguishes:

success

from:

no-data

and from:

error

A no-data result means the enrichment paths completed without producing a useful improvement.

It does not necessarily indicate a network or resolver failure.

Likewise, an error may be cached temporarily so RSSrch does not immediately repeat an unsuccessful request.

Enrichment requires usable external identity

Most network enrichment depends on an article URL or DOI.

An item with:

No URL
No DOI

normally cannot enter automatic metadata enrichment because RSSrch lacks a sufficiently useful external lookup identity.

This does not prevent the feed item itself from being used in Zotero or evaluated with whatever native metadata it already contains.

Enrichment does not guarantee complete metadata

Metadata Enrichment is opportunistic.

RSSrch cannot guarantee that every publisher exposes:

an abstract
complete creators
subject terms
a DOI
publication metadata

Some sources:

block automated requests
provide incomplete structured metadata
expose no useful external resolver
omit article-level subjects
return metadata that fails RSSrch’s identity or quality checks

In those cases, RSSrch retains the available feed metadata rather than inventing missing values.

Where enriched metadata appears

Enriched values can affect several parts of RSSrch.

Reader

The RSSrch Reader uses effective metadata and can show:

improved abstract
effective creators
publication information
date
DOI
Feed Subjects
enrichment provenance

A successful enrichment record is identified in the Metadata area as:

Enriched via [resolver]

Zotero feed item list

RSSrch can display effective enrichment-derived creator and date information in Zotero’s native feed-item columns without rewriting the underlying feed fields.

This improves the feed list while preserving Zotero’s native data record.

Rules and scoring

Applicable enriched metadata is available to RSSrch’s condition evaluation.

That means enrichment can improve the evidence available to:

Admission
Relevance Scoring
Action conditions

Feed Subjects

Accepted subject metadata can be applied as automatic Zotero tags and becomes available through the Feed Subjects field.

Native data and effective data are different concepts

It is useful to distinguish:

Native feed data

from:

Effective RSSrch data

For example:

Native abstract:
short publisher summary
 
Enrichment abstract:
full structured abstract
 
Effective RSSrch abstract:
full structured abstract

The native Zotero feed record can remain unchanged while RSSrch uses the better effective value for research processing and presentation.

This separation is central to the enrichment design.

Recommended workflow

For normal use:

Leave Enrich Feed Metadata + Subjects enabled.
Allow new feed items to be enriched automatically when they contain research-relevant metadata gaps.
Use the RSSrch Reader and rule diagnostics to inspect the resulting effective metadata.
Use Enrich Selected Feed Metadata Now when a specific paper still appears incomplete.
Use the single-item manual command when the strongest browser-translation fallback may be useful.
Use Enrich Feed Metadata + Subjects Now… when one feed contains many sparse items that should be processed.
Use Backfill Feed Subject Tags… only when subject-tag completion or cleanup is needed.
Keep Zotero automatic tags enabled if enrichment-derived Feed Subjects should be written as automatic tags.