Meta tags:
description=Redpanda Documentation: Guides, API references, and resources for event streaming.;

Headings (most frequently used words):

message, processing, batch, you, this, page, let, us, batching, performance, grouped, compatibility, policy, shrinking, batches, post, simple, online, edits, contribution, guide, components, highlight, hit, preview, attribute, title, what, do, think, of, know, more, contact, about, your, feedback, endpoint, example, on,

Text of the page (most frequently used words):
rpk (150), the (80), cloud (64), item (58), #preview (54), redpanda (51), api (45), cluster (40), and (38), html (34), state (31), with (30), return (30), configure (29), registry (29), data (27), batch (27), connect (26), for (25), context (24), create (24), this (23), security (23), const (22), product (21), list (21), you (21), messages (20), batching (19), message (19), components (18), schema (18), query (16), your (16), doc (15), use (15), hit (15), attribute (15), output (15), topic (15), profile (15), iceberg (15), connector (15), reference (14), index (14), topics (14), delete (14), algolia (13), that (13), are (13), processors (13), batches (13), sql (13), policy (12), aws (12), manage (12), https (11), www (11), com (11), title (11), storage (11), byoc (11), gcp (11), streaming (10), plane (10), set (10), objectid (10), processing (10), group (10), kafka (10), azure (10), server (9), map (9), when (9), window (9), private (9), configuration (9), config (9), into (9), count (9), input (9), part (9), multiple (9), describe (9), shadow (9), catalog (9), show (8), let (8), get (8), url (8), componenttitle (8), platform (8), parameters (8), time (8), type (8), has (8), add (8), open (8), breadcrumbs (8), snippet (8), intro (8), description (8), field (8), from (8), mcp (8), user (8), role (8), console (8), results (7), path (7), docs (7), overview (7), can (7), about (7), text (7), breadcrumb (7), highlight (7), before (7), source (7), resource (7), drop (7), compatibility (7), transform (7), alter (7), control (7), sink (7), not (6), new (6), all (6), link (6), version (6), sources (6), facet (6), selected (6), true (6), titles (6), length (6), search (6), algoliabtn (6), event (6), queryid (6), positions (6), status (6), such (6), pipeline (6), then (6), format (6), http (6), also (6), they (6), todo (6), broker (6), generate (6), catalogs (6), privatelink (6), mongodb (6), bloblang (6), method (5), origin (5), exact (5), match (5), plugin (5), hits (5), which (5), nestedhits (5), docindices (5), ask (5), stoppropagation (5), clickedobjectidsaftersearch (5), eventname (5), __autocomplete_indexname (5), __autocomplete_queryid (5), objectids (5), activeitemid (5), href (5), edit (5), down (5), reach (5), split (5), addresses (5), 100ms (5), period (5), aws_s3 (5), tar (5), parts (5), protocols (5), single (5), support (5), inputs (5), join (5), build (5), secret (5), mode (5), quotas (5), auth (5), log (5), access (5), monitor (5), management (5), currentquery (4), view (4), managed (4), home (4), see (4), filters (4), our (4), _tags (4), content (4), autocomplete (4), null (4), none (4), clickanalytics (4), while (4), non (4), matchingheading (4), tolowercase (4), within (4), across (4), document (4), function (4), click (4), feedback (4), work (4), complete (4), over (4), archive (4), allows (4), processor (4), 9092 (4), check (4), number (4), able (4), level (4), throughput (4), update (4), acl (4), install (4), txn (4), producers (4), mount (4), transforms (4), properties (4), dedicated (4), regions (4), serverless (4), billing (4), iam (4), service (4), peering (4), connection (4), centralized (4), egress (4), enable (4), quickstart (4), switch (4), nats_kv (4), byovpc (4), example (3), hasendpoint (3), endpointline (3), curlexample (3), catch (3), self (3), agentic (3), fallback (3), getitems (3), header (3), right (3), render (3), libraries (3), getalgoliafacets (3), queries (3), params (3), phrase (3), false (3), getalgoliaresults (3), order (3), items (3), foreach (3), _highlightresult (3), matchedindex (3), matchedtopleveltitle (3), renderproduct (3), seen (3), supported_preview_types (3), includes (3), its (3), without (3), btn (3), addeventlistener (3), may (3), any (3), slice (3), know (3), select (3), target (3), page (3), requires (3), windowed (3), contribution (3), read (3), will (3), flushed (3), them (3), sending (3), http_client (3), batched (3), conditions (3), benthos_stream (3), either (3), since (3), last (3), exceeds (3), metadata (3), means (3), how (3), decompress (3), consume (3), foo (3), individual (3), therefore (3), more (3), for_each (3), benefit (3), specify (3), performance (3), functions (3), deploy (3), failover (3), subject (3), globals (3), connections (3), tiers (3), timestamp (3), distinct (3), table (3), commits (3), secrets (3), policies (3), gbac (3), rbac (3), authorization (3), gateway (3), shadowing (3), migrate (3), snowflake (3), introduction (3), debezium (3), mirrormaker2 (3), jdbc (3), connectors (3), memory (3), redis (3), gcp_cloud_storage (3), sql_raw (3), azure_cosmosdb (3), pipelines (3), endpoint (2), enrichment (2), try (2), integrate (2), current (2), askaibutton (2), viewalllink (2), strict (2), typo (2), exacttoggle (2), define (2), values (2), tags (2), tagsbyfacet (2), tag (2), sourceid (2), checkbox (2), param (2), searchclient (2), indexname (2), test_redpanda (2), facetquery (2), maxfacethits (2), transformresponse (2), facethits (2), union (2), label (2), templates (2), term (2), optional (2), matching (2), __algoliatoggleexactmatch (2), exactmatch (2), exactparams (2), hitsperpage (2), attributestosnippet (2), snippetellipsistext (2), tagfilters (2), place (2), nondocitems (2), filter (2), unixtimestamp (2), infinity (2), nondocindex (2), full (2), clear (2), stale (2), array (2), navigate (2), otherwise (2), atag (2), setcontext (2), ctx (2), other (2), dropdown (2), references (2), autocompleteinstance (2), button (2), action (2), var (2), queryselector (2), detachedsearchbutton (2), key (2), should (2), _blank (2), rel (2), noopener (2), noreferrer (2), openchatwithquery (2), thanks (2), thumb_down (2), thumb_up (2), was (2), helpful (2), prefer (2), guide (2), issue (2), github (2), simple (2), back (2), have (2), post (2), lines (2), aggregate (2), here (2), sizes (2), but (2), break (2), send (2), passed (2), combine (2), value (2), zero (2), total (2), matches (2), size (2), byte_size (2), component (2), supports (2), mechanism (2), following (2), bucket (2), algorithm (2), gzip (2), unarchive (2), limit (2), protocol (2), select_parts (2), broken (2), using (2), behavior (2), only (2), behaviors (2), happen (2), sequentially (2), doesn (2), out (2), group_by_value (2), meta (2), kafka_partition (2), compress (2), operations (2), archiving (2), grouping (2), error (2), some (2), grouped (2), works (2), sometimes (2), case (2), instead (2), block (2), however (2), copy (2), metrics (2), produce (2), print (2), uninstall (2), offset (2), mountable (2), logdirs (2), proxy (2), apis (2), transactions (2), system (2), tables (2), bitwise (2), shift (2), left (2), operators (2), sum (2), min (2), max (2), bool_or (2), bool_and (2), avg (2), to_char (2), json (2), zone (2), numeric (2), pay (2), methods (2), based (2), authentication (2), hub (2), errors (2), troubleshoot (2), client (2), audit (2), maintenance (2), cli (2), postgresql (2), google (2), handling (2), authenticate (2), started (2), jira (2), migrator (2), streams (2), ingest (2), real (2), amazon (2), open_telemetry_collector (2), avro (2), noop (2), aws_dynamodb (2), sync_response (2), sql_select (2), sql_insert (2), retry (2), qdrant (2), gcp_bigquery_select (2), cache (2), timeplus (2), sftp (2), schema_registry (2), redpanda_migrator (2), redpanda_common (2), redis_streams (2), redis_pubsub (2), redis_list (2), otlp_http (2), otlp_grpc (2), nats_jetstream (2), nats (2), mqtt (2), kafka_franz (2), inproc (2), gcp_pubsub (2), azure_table_storage (2), azure_queue_storage (2), azure_blob_storage (2), aws_sqs (2), aws_kinesis (2), amqp_0_9 (2), what (2), documentation (2), labs (2), txt (2), tiny, reduce, clicks, pure, required, touppercase, curl, assistant, compose, respects, productparams, shows, products, 20data, 20plane, toggle, forces, layout, side, footer, tips, root, used, actual, core, concepts, getsources, groupby, tagsplugin, fetch, populate, menu, raw, facets, known, happens, require, every, spelling, word, note, deliberately, wrap, quotes, searchable, attributes, unordered, disables, proximity, quoted, returns, nothing, give, reliable, __algoliaexactmatch, typotolerance, removewordsifnoresults, optionalwords, maptoalgoliafilters, mixed, stay, sorted, compute, deep, best, sort, findindex, matchlevel, noresults, don, helper, deduped, badges, isarray, else, previewable, panel, tab, getitemurl, onactive, expose, instance, globally, custom, drive, capturing, renders, intentionally, namespaced, notes, below, __algoliaautocompleteinstance, wire, sidebar, modal, clicking, hidden, detached, queryselectorall, keydown, enter, preventdefault, domcontentloaded, initalgolia, answers, imprecise, always, verify, believe, close, esc, image, analytics, start, children, checked, typeof, setisopen, cookie, preferences, trust, terms, privacy, 2026, inc, rights, reserved, make, group_add, share, mail, community, extensive, updates, locally, something, want, change, changes, fixing, directly, online, edits, top, during, shutdown, remaining, waiting, above, merge, line, delimited, easier, parse, than, default, would, been, rfc1342, localhost, 4195, apply, each, good, compatible, affected, limitation, consider, breaking, configured, circumvent, specified, resulting, aren, expecting, capability, empty, added, causes, bytes, disregarding, exactly, routed, associated, considered, downstream, met, 5mb, 5_000_000, useful, creates, too, large, arrive, three, third, smaller, shrinking, easily, matched, concept, interchangeable, per, one, two, depending, collection, sent, falls, requests, write, payloads, travelling, through, represented, standard, examples, files, timestamp_unix_nano, there, vast, specialise, according, compresses, separate, archives, at_least_once, max_loops, errored, wipe, previous, attempt, until, success, executed, once, whole, avoid, same, brokers, bar, 500ms, already, wide, behave, way, documented, benthos_input_stream, small, must, consumed, partition, specifying, most, users, improving, background, benefits, like, formed, configuring, mixing, multi, sinks, conflicting, sacrificing, strong, delivery, guarantees, perform, powerful, reduction, serves, purposes, code, plain, copied, markdown, options, checking, glossary, logs, init, trim, prefix, partitions, unassign, assign, reset, rename, prompt, help, seek, shell, completion, grafana, dashboard, app, unmount, cancel, import, info, stdio, logout, login, token, commands, javascript, rust, golang, sdks, object, virtual, information, pg_user, pg_type, pg_statio_user_tables, pg_shadow, pg_settings, pg_roles, pg_proc, pg_namespace, pg_language, pg_index, pg_description, pg_depend, pg_database, pg_constraint, pg_class, pg_authid, pg_attribute, pg_attrdef, row_number, rank, percent_rank, ntile, nth_value, lead, last_value, lag, first_value, dense_rank, cume_dist, unix_seconds, unix_millis, unix_micros, to_timestamp, timestamp_trunc, timestamp_seconds, timestamp_millis, timestamp_micros, format_timestamp, extract, date_trunc, current_timestamp, upper, substring, strpos, starts_with, replace, position, lower, ends_with, concat, regexp_replace, regexp_match, posix, regular, expressions, regex, string, shobj_description, pg_typeof, pg_total_relation_size, pg_table_size, pg_table_is_visible, pg_size_pretty, pg_relation_is_publishable, pg_get_userbyid, pg_get_statisticsobjdef_columns, pg_get_indexdef, pg_get_expr, pg_get_constraintdef, pg_encoding_to_char, pg_backend_pid, obj_description, nullif, has_schema_privilege, generate_series, current_schema, current_database, col_description, coalesce, sqrt, sinh, sin, sign, round, random, power, least, greatest, floor, exp, cosh, ceil, cbrt, abs, math, json_extract_path_text, json_extract_path, json_array_length, json_array_extract, boolean, for_min, for_max, variance, var_samp, var_pop, stddev_samp, stddev_pop, stddev, regr_syy, regr_sxy, regr_sxx, regr_slope, regr_r2, regr_intercept, regr_count, regr_avgy, regr_avgx, covar_samp, covar_pop, corr, statistics, percentile_disc, percentile_cont, ordered, row, aliases, interval, geometry, geography, date, bytea, bool, types, where, having, outer, intersect, except, clauses, nodes, execs, revoke, grant, statements, schemas, comment, keywords, partner, integrations, notifications, activity, payment, safety, reliability, availability, encryption, lists, acls, global, transit, choose, cidr, ranges, network, design, ports, networking, terraform, provider, deprecation, codes, runbook, disaster, recovery, programmable, push, deserialization, validation, contexts, tune, lakehouse, databricks, unity, glue, rest, sample, logging, upgrades, profiles, organization, mysql, heartbeat, checkpoint, gcs, bigquery, sizing, disable, converters, serialization, test, paginate, follower, fetching, consumer, offsets, leader, pinning, idempotent, develop, degraded, nested, fields, enabled, editor, dapper, npgsql, php, pdo, java, psycopg2, oltp, olap, issues, retrieval, augmented, generation, rag, joining, filtering, sampling, workflows, dynamodb, cdc, patterns, cookbooks, unified, synchronous, responses, sensor, telemetry, aurora, roles, web, services, credentials, arithmetic, walkthrough, guides, logger, prometheus, gcp_cloudtrace, tracers, to_the_end, skip_bom, re_match, json_documents, json_array, csv, chunker, scanners, system_window, buffers, local, rate, limits, ttlru, ristretto, multilevel, memcached, lru, caches, xml, workflow, try_catch, text_chunker, string_split, sleep, slack_thread, schema_registry_encode, schema_registry_decode, redis_script, rate_limit, protobuf, parse_log, parquet_encode, parquet_decode, parallel, openai_translation, openai_transcription, openai_speech, openai_image_generation, openai_embeddings, openai_chat_completion, ollama_moderation, ollama_embeddings, ollama_chat, nats_request_reply, mutation, metric, mapping, json_schema, jmespath, insert_part, group_by, google_drive_search, google_drive_list_labels, google_drive_download, gcp_vertex_ai_embeddings, gcp_vertex_ai_chat, dedupe, cohere_rerank, cohere_embeddings, cohere_chat, cached, branch, bounds_check, benchmark, aws_lambda, aws_dynamodb_partiql, aws_bedrock_embeddings, aws_bedrock_chat, a2a_message, splunk_hec, snowflake_streaming, snowflake_put, slack_reaction, slack_post, salesforce_sink, reject_errored, reject, redis_hash, questdb, pinecone, opensearch, gcp_bigquery_write_api, gcp_bigquery, elasticsearch_v8, drop_on, cyborgdb, azure_data_lake_gen2, aws_sns, aws_kinesis_firehose, arc, outputs, splunk, spicedb_watch, slack_users, slack, sequence, salesforce_graphql, salesforce_cdc, salesforce, redis_scan, read_until, postgres_cdc, oracledb_cdc, mysql_cdc, mongodb_cdc, microsoft_sql_server_cdc, http_server, gcp_spanner_cdc, git, aws_dynamodb_cdc, aws_cloudwatch_logs, unit, testing, resources, process, paths, interpolation, contextual, variables, remote, replicas, byovnet, architecture, hundreds, deployments, run, own, infrastructure, fully, foundation, everything, agents, please, robot, submitting, submit, contact, think, beep, boop, bot, familiar, syntax, learn, specific, llms,

Text of the page (random words):
send batches when they reach 10 messages or when 100ms has passed since the last batch batching count 10 period 100ms however a small number of inputs such as kafka must be consumed sequentially in this case by partition and therefore benefit from specifying your batch policy at the input level instead input kafka addresses todo 9092 topics benthos_input_stream batching count 10 period 100ms output kafka addresses todo 9092 topic benthos_stream inputs that behave this way are documented as such and have a batching configuration block sometimes you may prefer to create your batches before processing in order to benefit from batch wide processing in which case if your input doesn t already support a batch policy you can instead use a broker which also allows you to combine inputs with a single batch policy input broker inputs resource foo resource bar batching count 50 period 500ms this also works the same with output brokers grouped message processing and some processors such as while are executed once across a whole batch you can avoid this behavior with the for_each processor pipeline processors for_each while at_least_once true max_loops 0 check errored processors catch wipe any previous error resource foo attempt this processor until success there s a vast number of processors that specialise in operations across batches such as grouping and archiving for example the following processors group a batch of messages according to a metadata field and compresses them into separate tar gz archives pipeline processors group_by_value value meta kafka_partition archive format tar compress algorithm gzip output aws_s3 bucket todo path docs meta kafka_partition count files timestamp_unix_nano tar gz for more examples of batched or windowed processing check out this document compatibility redpanda connect is able to read and write over protocols that support multiple part messages and all payloads travelling through redpanda connect are represented as a multiple part message therefore all components within redpanda connect are able to work with multiple parts in a message as standard when messages reach an output that doesn t support multiple parts the message is broken down into an individual message per part and then one of two behaviors happen depending on the output if the output supports batch sending messages then the collection of messages are sent as a single batch otherwise redpanda connect falls back to sending the messages sequentially in multiple individual requests this behavior means that not only can multiple part message protocols be easily matched with single part protocols but also the concept of multiple part messages and message batches are interchangeable within redpanda connect shrinking batches a message batch or multiple part message can be broken down into smaller batches using the split processor input consume messages that arrive in three parts resource foo processors drop the third part select_parts parts 0 1 then break our message parts into individual messages split size 1 this is also useful when your input source creates batches that are too large for your output protocol input aws_s3 bucket todo pipeline processors decompress algorithm gzip unarchive format tar limit batch sizes to 5mb split byte_size 5_000_000 batch policy when an input or output component has a config field batching that means it supports a batch policy this is a mechanism that allows you to configure exactly how your batching should work on messages before they are routed to the input or output it s associated with batches are considered complete and will be flushed downstream when either of the following conditions are met the byte_size field is non zero and the total size of the batch in bytes matches or exceeds it disregarding metadata the count field is non zero and the total number of messages in the batch matches or exceeds it a message added to the batch causes the check to return to true the period field is non empty and the time since the last batch exceeds its value this allows you to combine conditions output kafka addresses todo 9092 topic benthos_stream either send batches when they reach 10 messages or when 100ms has passed since the last batch batching count 10 period 100ms a batch policy has the capability to create batches but not to break them down if your configured pipeline is processing messages that are batched before they reach the batch policy then they may circumvent the conditions you ve specified here resulting in sizes you aren t expecting if you are affected by this limitation then consider breaking the batches down with a split processor before they reach the batch policy post batch processing a batch policy also has a field processors which allows you to define an optional list of processors to apply to each batch before it is flushed this is a good place to aggregate or archive the batch into a compatible format for an output output http_client url http localhost 4195 post batching count 10 processors archive format lines the above config will batch up messages and then merge them into a line delimited format before sending it over http this is an easier format to parse than the default which would have been rfc1342 during shutdown any remaining messages waiting for a batch to complete will be flushed down the pipeline back to top simple online edits for simple changes such as fixing a typo you can edit the content directly on github edit on github or open an issue to let us know about something that you want us to change open an issue contribution guide for extensive content updates or if you prefer to work locally read our contribution guide was this helpful thumb_up thumb_down group ask in the community mail share your feedback group_add make a contribution thanks for your feedback configuration windowed processing 2026 redpanda data inc all rights reserved privacy policy terms status trust cookie preferences for api endpoint results we show a tiny example to reduce clicks this is pure ui enrichment not required by algolia const hasendpoint preview preview method preview path let endpointline let curlexample if hasendpoint const method preview method get touppercase const path preview path endpointline method path let origin try origin new url preview url origin catch e no op curlexample curl x method origin origin path path integrate ai assistant with the current query const askaibutton currentquery html if typeof window openchatwithquery function window openchatwithquery currentquery true autocompleteinstance setisopen false ask ai about currentquery compose a view all results link that respects product version context const componenttitle cloud const productparams if componenttitle cloud return product 0 cloud product 1 connect if componenttitle streaming return product 0 streaming product 1 connect if componenttitle connect return product 0 connect if componenttitle data platform return product 0 cloud product 1 streaming product 2 connect if componenttitle self managed return product 0 streaming product 1 connect if componenttitle home return home shows all products if componenttitle agentic data plane return product 0 agentic 20data 20plane return const viewalllink currentquery html view all results exact match toggle forces a strict no typo no fallback query see getitems const exacttoggle html window __algoliatoggleexactmatch e target checked exact match layout header filters ai results list right side preview footer tips render html dropdown exacttoggle askaibutton children preview html preview breadcrumbs html preview breadcrumbs map breadcrumb html event stoppropagation click analytics requires clickanalytics true and a queryid docs https www algolia com doc api reference api parameters clickanalytics aa clickedobjectidsaftersearch eventname preview selected index state context preview __autocomplete_indexname queryid state context preview __autocomplete_queryid objectids state context preview objectid positions state activeitemid 1 positions start at 1 href breadcrumb u breadcrumb t components highlight hit preview attribute title preview intro components highlight hit preview attribute intro preview description components highlight hit preview attribute description preview api html preview api hasendpoint html endpoint endpointline example curlexample preview image html supported_preview_types includes preview type tolowercase html event stoppropagation aa clickedobjectidsaftersearch eventname preview selected index state context preview __autocomplete_indexname queryid state context preview __autocomplete_queryid objectids state context preview objectid positions state activeitemid 1 href preview url preview objectid html event stoppropagation aa clickedobjectidsaftersearch eventname preview selected index state context preview __autocomplete_indexname queryid state context preview __autocomplete_queryid objectids state context preview objectid positions state activeitemid 1 target _blank rel noopener noreferrer href preview url preview objectid preview titles preview titles length 0 html on this page preview titles slice 0 15 map title html supported_preview_types includes state context preview type tolowercase html event stoppropagation aa clickedobjectidsaftersearch eventname preview selected index state context preview __autocomplete_indexname queryid state context preview __autocomplete_queryid objectids state context preview objectid positions state activeitemid 1 href preview url preview objectid title h components highlight hit title attribute t html event stoppropagation aa clickedobjectidsaftersearch eventname preview selected index state context preview __autocomplete_indexname queryid state context preview __autocomplete_queryid objectids state context preview objectid positions state activeitemid 1 target _blank rel noopener noreferrer href preview url preview objectid title h components highlight hit title attribute t to select to navigate esc to close viewalllink root define our data sources filters facet values for the _tags facet used by the tags plugin docs actual content hits sources overview https www algolia com doc ui libraries autocomplete core concepts sources getsources query state const tagsbyfacet groupby state context tagsplugin tags tag tag facet return sourceid filters getitems fetch facet values for _tags to populate the checkbox menu getalgoliafacets https www algolia com doc ui libraries autocomplete api reference sources param getalgoliafacets return getalgoliafacets searchclient queries indexname test_redpanda facet _tags params facetquery https www algolia com doc api reference api parameters facetquery maxfacethits 50 https www algolia com doc api reference api parameters maxfacethits transformresponse facethits we return raw server facets union with known selected happens in render const server facethits 0 map h label h label facet _tags return server templates header return null item return null sourceid docs getitems when exact match is on require every term exact spelling and no optional word fallback note we deliberately do not wrap the query in quotes for a phrase match the index s searchable attributes use unordered which disables phrase proximity matching so a quoted phrase returns nothing strict params give reliable exact term matching see window __algoliatoggleexactmatch const exactmatch window __algoliaexactmatch const exactparams exactmatch typotolerance false removewordsifnoresults none optionalwords getalgoliaresults https www algolia com doc ui libraries autocomplete api reference sources param getalgoliaresults return getalgoliaresults searchclient queries indexname test_redpanda query params clickanalytics true https www algolia com doc api reference api parameters clickanalytics hitsperpage 10 https www algolia com doc api reference api parameters hitsperpage attributestosnippet 25 https www algolia com doc api reference api parameters attributestosnippet snippetellipsistext https www algolia com doc api reference api parameters snippetellipsistext tagfilters maptoalgoliafilters tagsbyfacet https www algolia com doc api reference api parameters tagfilters exactparams transformresponse hits re order mixed hits so docs stay in place while non doc items are time sorted and compute a matchingheading to deep link to the best title match return hits map nestedhits const docindices new map nestedhits foreach item index if item type doc docindices set index item const nondocitems nestedhits filter item item type doc sort a b b unixtimestamp infinity a unixtimestamp infinity let nondocindex 0 docindices foreach item index if item _highlightresult item _highlightresult titles const matchedindex item _highlightresult titles findindex title title t matchlevel full if matchedindex 1 const matchedtopleveltitle item titles matchedindex if matchedtopleveltitle item matchingheading matchedtopleveltitle h nestedhits index item return nestedhits map item index docindices has index docindices get index nondocitems nondocindex templates noresults state html clear preview so we don t show stale content for a query with no hits state context preview null if state query return return html no results for state query believe this query should return results let us know header items html if items length return return html results item item components html const matchingheading item matchingheading helper render product as deduped badges const renderproduct product if array isarray product const seen new set return product filter p const k p tolowercase if seen has k return false seen add k return true map p html p else if product return html product return null if previewable navigate within the panel otherwise open a new tab const atag supported_preview_types includes item type tolowercase html components highlight hit item attribute title item text html item text components snippet hit item attribute text item intro components snippet hit item attribute intro item description components snippet hit item attribute description html item intro components snippet hit item attribute intro item description components snippet hit item attribute description item type renderproduct item product item api html item api item version html item version html components highlight hit item attribute title item breadcrumbs html item breadcrumbs length 2 item breadcrumbs slice 1 item breadcrumbs length 1 map breadcrumb html breadcrumb t item breadcrumbs length 2 item breadcrumbs slice 1 map breadcrumb html breadcrumb t item text components snippet hit item attribute text item intro components snippet hit item attribute intro item description components snippet hit item attribute description item type renderproduct item product item version html item version return html atag getitemurl item return item url item objectid onactive item setcontext state const ctx state context setcontext ctx preview item expose the autocomplete instance globally so other ui our cu...