Step 5: Advanced ship tags

Preserving author order to reliably identify primary and secondary ships

The short ship column in step 4 reads directly from #ship — which works, but has a limitation. Calibre’s tag columns sort alphabetically, so for fics with multiple pairings the “first” ship in your column is whoever’s name comes first in the alphabet, not who the author foregrounded. A Johnlock fic where the author listed Mycroft/Greg too might show it as the main ship because it comes first in the Calibre relationship tags.

If that’s good enough for your library (if your fics tend to have only one ship at a time), step 4 is fine as-is. This step sets up a more robust pipeline that captures ships in author order and extracts them into dedicated columns.

I’ve set it up to capture two ships: the first one in order of appearance I set as primary, the second one as secondary. You could in theory expend this to however many, or remove the secondary and just have one.

What you’ll end up with

Two new Calibre columns — #primary_slash and #secondary_slash — each containing a single ship in the order the author listed them. These feed a more accurate short ship template, and because they’re tag columns, clicking a ship name in the tag browser shows every fic where that’s the primary pairing.

A better #short_ships column that combines the primary and secondary ships in short format. Ready to drop into your covers

How it works

The key problem is timing: by the time data reaches Calibre, sorting has already happened. The solution is to intercept the relationship data inside FFF’s processing pipeline, before Calibre touches it, and extract the ships we want into separate variables.

The pipeline does five things in order:

1. Creates three internal FFF scratch variables: all_slashes, primary_slash, secondary_slash
2. Copies the raw ships data — still in author order — into all_slashes
3. Filters all_slashes down to slash-only relationships, dropping all & platonic tags
4. Extracts the first ship into primary_slash and the second into secondary_slash
5. Writes both to Calibre columns

Create the Calibre columns

Before editing personal.ini, create the two columns in Calibre.

In Preferences → Add your own columns, add:

Single-value tag columns don’t sort in any meaningful way, so you get the tag browser benefit — clickable ship names — without any sorting side effects.

Update personal.ini

Add the following to your [defaults] section. The order matters — each block depends on the one above it.

# Create internal scratch variables for the pipeline
add_to_extra_valid_entries: ,all_slashes,primary_slash,secondary_slash

# Copy ships into all_slashes while still in author order
 include_in_all_slashes: ships

# Tell FFF not to alphabetise these fields during processing
 keep_in_order_ships:true
 keep_in_order_all_slashes:true

# Filter all_slashes down to slash-only relationships
 include_metadata_pre:
    all_slashes=~/

# Copy the filtered list into both extraction variables
 include_in_primary_slash: all_slashes
 include_in_secondary_slash: all_slashes

Check if you have an add_to_replace_metadata block. If you do skip the first line:

add_to_replace_metadata:
# Extract first ship only into primary_slash
 primary_slash_LIST=>^([^,]+).*$=>\1

# Extract second ship only into secondary_slash
# (clears the field if there is no second ship)
 secondary_slash_LIST=>^[^,]+$=>
 secondary_slash_LIST=>^[^,]+,\s*([^,]+).*$=>\1

And add the column mappings to your [archiveofourown.org] section:

add_to_custom_columns_settings:
    primary_slash=>#primary_slash
    secondary_slash=>#secondary_slash

Update the short ship template

Replace your step 4 short ship template with this version, which reads from the new columns and falls back to #ship for fics downloaded before the pipeline was set up — or for fics that have since been deleted from AO3.

program:
# Read from the dedicated columns
s0 = field('#primary_slash');
s1 = field('#secondary_slash');
result = '';

# Fallback: if both columns are empty, extract from #ship directly
if !s0 && !s1 then
    ships_raw = field('#ship');
    if contains(ships_raw, '/', '1', '') then
        fallback = re(ships_raw, '^(?:[^/,]+,\s*)*([^,]+/[^,]+)(?:,.*)?$', '\1');
        if contains(fallback, '/', '1', '') && !contains(fallback, ',', '1', '') then
            s0 = fallback
        fi
    fi
fi;

# Translate slot 1
t1 = '';
if s0 then
    s1_lc = lowercase(s0);
    if contains(s1_lc, 'sherlock holmes/john watson|john/sherlock|sherlock/john', '1', '') then t1 = 'Johnlock'
    elif contains(s1_lc, 'shane hollander/ilya rozanov', '1', '') then t1 = 'Hollanov'
    # Add your ships here
    else t1 = re(s0, '(?i)([^/]+)/([^/]+)', '\1/\2')
    fi
fi;

# Translate slot 2
t2 = '';
if s1 then
    s2_lc = lowercase(s1);
    if contains(s2_lc, 'sherlock holmes/john watson|john/sherlock|sherlock/john', '1', '') then t2 = 'Johnlock'
    elif contains(s2_lc, 'shane hollander/ilya rozanov', '1', '') then t2 = 'Hollanov'
    # Add your ships here
    else t2 = re(s1, '(?i)([^/]+)/([^/]+)', '\1/\2')
    fi
fi;

# Combine, skip duplicates
if t1 then result = t1 fi;
if t2 then
    if !result then result = t2
    elif lowercase(t2) != lowercase(result) then result = result & ', ' & t2
    fi
fi;

if !result then return 'Gen' else return result fi

The ship translation table works exactly the same way as in step 4 — add your ships to both the slot 1 and slot 2 sections following the same pattern.

Using primary ship for cover selection

With #primary_slash reliably populated, you can add per-ship rules to generate_cover_settings. Ship-specific rules should sit above fandom rules, since the first match wins:

generate_cover_settings:
    ${primary_slash} => Mycroft Holmes/Greg Lestrade => Classics Mystrade
    ${category} => [Ss]herlock => Classics Sherlock
    ...

This is how you get a dedicated cover for a specific pairing, because your OTP deserves its own template. This also works across fandoms.