Warning: Modifying these configurations may cause application instability or unexpected behavior. Please ensure you fully understand the impact before making any changes. Additionally, you can use the Configuration applies to dropdown to determine the scope of your changes (e.g., applying them to “All localized catalogs”) before saving.

image

Once configurations are done, Button Save needs to be pressed so that settings are effective.

General

image

General

Parameter Default value Description
Media base url https://master.gally.dev/media/catalog/product/ The base URL path used by the application to locate, resolve, and serve media assets, such as product images.
Gally back office url https://master.gally.dev The primary root URL used to access the Gally back office administrative interface.

Email Settings

Parameter Default value Description
Default sender name Gally Master The default display name that will appear as the sender when the system dispatches automated emails to users or customers.
Default sender email noreply+master@gally.dev The default email address utilized by the system to send automated communications and notifications.

Index

These configurations allows you to manage the settings related to search indices, state management, and templates.

Note: Some of the descriptions below regarding standard indexing parameters (such as shards, replicas, and prefixes) are drawn from standard search engine knowledge outside of your provided sources. You may want to independently verify these descriptions align with your exact system behavior.

General

Parameter Default value Description
Index prefix gally The standard prefix applied to the names of all indices generated by the application.
Timestamp pattern _ The date and time format used to append a timestamp to the index name, ensuring uniqueness for each generated index.
Number of shards 1 The number of primary shards allocated to the index for data distribution.
Number of replicas 0 The number of replica copies of the shards allocated for the index.
Ghost index timeout (s) 3600 Gally derelict indices resulting from a failed full reindex are considered ghost after this amount of time (in seconds) has elapsed since their creation. You can reduce this amount of time to speed up ghost indices cleanup, but take care to add a safety on top of the maximum reindexing duration of the more complex index of your platform (usually a catalog_product/product search index). Defaults to 172,800 seconds (2 days), minimum value: 3600 (1 hour).

Index State Management

Parameter Default value Description
ISM prefix gally The standard prefix used for naming Index State Management (ISM) policies.
Rollover after (days) 30 The number of days after which an active index will automatically roll over and transition.
Delete after (days) 365 The retention period for your indices. Older data will be deleted daily after reaching this threshold.

Index Template

Parameter Default value Description
Index template prefix gally The prefix applied to index templates used by the application to configure new indices automatically.

The Configuration menu in Gally includes the Search tab, which handles the core settings for search pipelines, spatial searches, sorting behavior, and autocomplete limits.

Note: As with previous sections, standard descriptions for these parameters are provided based on standard search engine and e-commerce platform behavior.

General

Parameter Default value Description
Pipeline prefix gally-llm-pipeline- The standard prefix applied to data processing or LLM pipelines used by the search engine.
Default price group ID 0 The fallback customer or price group ID used to index and display pricing when no specific group is provided.
Default reference location 48.981299, 2.309959 The default geographical coordinates (latitude, longitude) used as a fallback center point for geo-spatial searches or distance calculations.

Search

Parameter Default value Description
Default date field format yyyy The default format pattern used by the search engine for parsing and indexing date fields.
Default distance unit km The default unit of measurement (kilometers) applied to geo-distance queries and geographical facets.
Default date facet interval 1y The default time interval (e.g., 1 year) used when aggregating and grouping date fields into search facets.
Default geo-distance ranges [object Object],... A predefined array of distance ranges used by default to build geographical facets (e.g., 0-5km, 5-10km) for the front-end users.
Code with default descending sort _count,_score A comma-separated list of sort codes or metrics (such as relevance score or hit count) that are automatically sorted in descending order by default.

Autocomplete

Parameter Default value Description
Max attr. options autocomplete 3 The maximum number of specific attribute options (e.g., “Red”, “Blue”) displayed within the autocomplete suggestions dropdown.
Max attr. autocomplete 3 The maximum number of distinct attributes (e.g., “Color”, “Size”) suggested in the autocomplete dropdown when a user types a query.

Relevance

The Relevance configuration section in Gally allows you to fine-tune the core search engine behaviors, adjusting how aggressively the engine matches user queries with product data and how it calculates the final score (ranking) of the results.

These settings control the base full-text matching logic for user queries.

Configuration Field Default Value Description
Minimum should match 100% Defines the minimum percentage of words from the search query that a product must contain to be considered a valid result. A value of 100% ensures highly relevant results by requiring all words to be present.
Tie breaker 1 Determines how much the scores of the lesser-matching fields affect the final score when a search term is found in multiple attributes.

These settings allow you to reward products that contain the exact phrase searched by the user (the exact words in the exact same order).

Configuration Field Default Value Description
Enable phrase boost No Activates phrase-level boosting to reward products that contain the exact phrase searched by the user.
Boost value 0 The multiplier applied to the score when the exact phrase is matched.

Stopwords Detection

This section dynamically ignores extremely common words to improve performance and relevance.

Configuration Field Default Value Description
Cutoff frequency 0,15 Determines the frequency threshold (e.g., 15%) at which a term is dynamically considered a stopword and ignored by the engine because it appears in too many products.

Fuzzy search helps catch typos and spelling mistakes made by users.

Configuration Field Default Value Description
Enable fuzzy search Yes Activates typo tolerance to help catch spelling mistakes made by users.
Fuzzy search similarity AUTO The edit distance allowed for typo correction, usually dynamically adjusted based on the length of the searched word.
Fuzzy search prefix 1 The number of initial characters that must match exactly before fuzzy logic is applied (e.g., a prefix of 1 means the first letter must be strictly identical).
Maximum number of expansions 10 The maximum number of alternate typo-corrected terms the engine will generate and search for.

Allows the engine to match words that sound similar even if they are spelled differently.

Configuration Field Default Value Description
Enable phonetic search Yes Activates phonetic text analysis, allowing the engine to match words that sound similar even if they are spelled differently.

Rewards products where the searched terms appear close to each other or at the very beginning of the field.

Configuration Field Default Value Description
Span boost 10 The multiplier applied to the score when terms are found in close proximity to each other or at the very beginning of the field.
Span slop 10 The maximum number of intervening words allowed between the searched terms for them to still receive the proximity boost.
Span in order No Determines whether the words must appear in the exact same order as typed by the user to receive the boost.

Boost

The Boost configuration section in Gally allows you to fine-tune the core boost behaviors.

Model: attribute value

Allows the admin to list the type of attribute that can be used in proportionnal to an attribute value boost.

Configuration Field Default Value Description
Proportional boost attribute types Int, Float Specifies the data types of the attributes (such as integers or floats) that are eligible to be used for proportional boosting.

Thesaurus

These settings allow you to configure the behavior of the thesaurus, controlling how synonyms and expansions are applied to search queries to provide relevant alternative results.

General

This section defines the global limits and behavior of the thesaurus rewriting process.

Configuration Field Default Value Description
Maximum number of rewrites 2 Maximum number of thesaurus rules applied at a given time to a given search query to produce alternative queries. That number applies first to the synonyms rules and then the expansion rules. For instance if the setting’s value is 2, it means each alternative query will be the result of the application of at most 2 synonyms rules and at most 2 expansion rules. But if you have 10 synonym rules and 5 expansion rules, they could all end up being applied by pairs. So be careful about augmenting this setting’s value, especially if you already have a lot of rules with long lists of alternative terms.

Synonym

These settings control the activation and scoring impact of synonym rules.

Configuration Field Default Value Description
Enable synonyms Yes Determines whether synonym rules are actively applied to search queries.
Synonym weight divider 10 The factor by which the score of a synonym match is divided. This ensures that exact matches generally rank higher than synonym matches.

Expansion

These settings control the activation and scoring impact of expansion rules.

Configuration Field Default Value Description
Enable expansions Yes Determines whether expansion rules are actively applied to search queries.
Expansion weight divider 10 The factor by which the score of an expansion match is divided, adjusting its relevance score relative to the original search term.

These settings allow you to configure the behavior of vector search (often referred to as neural, k-NN, or semantic search). Vector search leverages machine learning models to understand the broader context and meaning of a user’s query, allowing the engine to return highly relevant products based on conceptual similarity rather than relying strictly on exact keyword matching.

Search Configuration

Configuration Field Default Value Description
K value 100 Determines the number of nearest neighbors (top matching documents) the engine retrieves from the cluster during the initial vector search phase. A higher value increases the number of potential matches (recall) but may slightly impact performance.
Minimum score 0,015 Sets the absolute similarity threshold required for a vector match to be considered valid. Results with a similarity score below this value will be considered irrelevant and automatically excluded from the final results.
Use for spellcheck No Determines whether vector search should be utilized to retrieve or blend results when the system detects a misspelled query or applies a typo correction.
Use for exact match No Defines whether vector search should be executed and combined with standard full-text results even when the engine has already successfully found exact keyword matches.
Use on no result No Determines if the engine should automatically fall back to vector search to provide conceptual recommendations when the standard full-text query returns 0 results, helping to prevent “dead-end” empty pages.

Explain

These settings allow you to configure the behavior of the Explain feature, which provides detailed insights into why specific products are returned for a given search query and how their relevance scores are calculated.

General

Configuration Field Default Value Description
Highlight collector fields yes Determines whether to visually highlight the matched search terms within the aggregate collector fields when analyzing search results. This helps administrators quickly identify exactly which tokens or aggregated attributes contributed to a product’s match and final score.

Term suggestion

Term Suggestion

These settings allow you to configure the behavior of term suggestions. By leveraging historical search data, you can dynamically guide users toward popular and relevant queries based on past browsing activity.

General

Configuration Field Default Value Description
Suggested terms period (in day) 180 Determines the historical timeframe (in days) from which the system will pull past user search data to generate suggestions. For example, a value of 180 means suggestions are based on the last 6 months of search activity.
Boost suggested terms period (in day) 7 Applies a relevance boost to terms that have been frequently searched within a more recent, shorter timeframe (e.g., the last 7 days). This helps trending or highly active recent searches appear higher in the suggestions list.
Max suggestion terms 10 The maximum number of suggested search terms that will be returned and displayed to the user at any given time.
This site uses Just the Docs, a documentation theme for Jekyll.