FME Hub Alerts

FME Hub user antoine just uploaded a new template to the FME Hub.FME-SQL-LLM Report GeneratorType: Template workspace (Form)Purpose: Turn a plain-language question about a table into a planned set of SQL queries, execute them in DuckDB, and produce a Markdown → HTML report using three lightweight LLM roles (Planner, SQL Generator, Analyst).What it doesStandardizes input: reads any tabular file (Excel/CSV/Parquet), normalizes headers, and writes a temporary Parquet dataset.Profiles data: extracts schema, summary stats, and sample rows with DuckDB (compact “metadata” for prompts).Plans the analysis (Planner LLM): converts the goal + metadata into formal sub-questions and SQL intents.Drafts SQL (SQL Generator LLM): returns executable SQL per sub-question (minimal JSON: just sql).Executes safely in DuckDB: runs each query with guardrails (e.g., COPY (SELECT ...), semicolon stripping, optional LIMIT cap).Analyzes results (Analyst LLM): synthesizes named result sets into a Markdown narrative

FME Hub user andreas_h just uploaded a new transformer to the FME Hub.URLResolverOverviewThe URLResolver is an Custom Transformer that resolves relative URLs against a base URL with input validation. It ensures base URLs are absolute and relative URLs are relative before combining them using urllib.parse.urljoin,Key FeaturesURL validation (base_url must be absolute, relative_url must be relative)RFC 3986 compliant URL resolution using urllib.parse.urljoinHandles all relative URL types (root-relative, parent-relative, document-relative, query-only, fragment-only)Automatic path navigation resolution (.. and . segments)Error reporting with detailed messagesInput AttributesThe transformer expects the following attributes on input features:AttributeRequiredValidationDescriptionExample_base_urlYesMust be absolute with schemeThe base URL (must include protocol)https://example.com/path/page.html_relative_urlYesMust be relative without schemeThe relative URL to resolve../other.html, /path/file.

FME Hub user takashi just uploaded a new format to the FME Hub.Reads DEM data file(s) (GML format) of the Japanese Fundamental Geospatial Data (FGD) as raster feature(s).This reader generates one raster feature per one source data file; the feature type name will always be set to "DEMRASTER".You can read any type DEM data (DEM1A, DEM5A, DEM5B, DEM10A, or DEM10B) with this reader, but DO NOT change source data file names since it determines the DEM type based on the naming rule for FGD DEM data files.This reader skips reading a source file when the schema location or the coordinate system defined in the source file is invalid, or the source file name is invalid as an FGD DEM data file. If you don't want to validate schema location, set "No" to the "Validate Schema Location" parameter. This reader considers these srsName as valid, and sets a corresponding coordinate system (EPSG code) to the output raster feature. Note: at 2025-10-25, since EPSG code for JGD2024 isn't defined, sets EPSG:

FME Hub user mark2atsafe just uploaded a new transformer to the FME Hub.Creates a datetime interval using values entered via published parameters.Would you like to know more? Click here to find out more details!

FME Hub user antoine just uploaded a new template to the FME Hub.Translate Word Docs (EN→FI) with FME + Poro (Ollama)Category: AI / NLP • Type: Workspace Template (.fmwt)FME Version: 2024.2+Tags: translation, Word, DOCX, Finnish, LLM, Ollama, PoroOverviewThis template translates Microsoft Word (.docx) documents from English to Finnish using the FME Word Reader/Writer and a small Finnish LLM (Llama-Poro-2-8B-Instruct) served locally via Ollama.It keeps element order and layout by translating text-bearing features in place, then writing a new Finnish DOCX.Why this template?- No XML wrangling—works directly with the Word Reader/Writer.- Runs locally (good for sensitive docs).- Chunked context per block for more consistent translations.When to UseYou have DOCX files with standard paragraphs, lists, tables, captions.You want local inference and decent coherence without cloud services.You don’t need enterprise-scale throughput (see Notes & Limits).FeaturesWord Reader → filter text elemen

FME Hub user antoine just uploaded a new template to the FME Hub.ExprAgent_Ollama Generate & Run FME Expressions with an LLMType: Template workspace (Form)Purpose: Convert a plain-English task into an Arithmetic Editor expression using a local LLM (Ollama), then evaluate that expression on every feature and write the numeric result to _result.What it doesScans the input schema and builds a compact JSON description of attributes.Calls an Ollama tools-enabled model via /api/chat with a single tool (eval_expression) so the model returns only an expression.Cleans the returned textEvaluates the expression for each feature and stores the value in _result.Why it’s usefulTurn a request like:add 5 to _creation_instance and multiply by 3check if the id of my instance is eveninto a valid expression:(@Value(_creation_instance) + 5) * 3…then apply it across your data—no manual expression authoring needed.Inputs & outputsInput: Any feature type with attributes referenced by your task.Output

FME Hub user vhruska just uploaded a new transformer to the FME Hub.Coda Table ReaderThis transformer lets you read data from a Coda table into FME.👉 Generate your Coda API key hereHow It WorksDoc Id: The unique ID of your Coda document.Table Id: The unique ID of the table inside that document.API Token: Your personal Coda API key for authentication.OutputReads rows from the specified Coda table. Optionally outputs the column schema (table structure) if you enable it in settings.Because Coda tables can have dynamic structures, you will need to use an AttributeExposer transformer after this to make sure all fields are visible and usable in your workflow.Use CasesPull Coda data into FME workflows for analysis, transformation, or integration. Retrieve both data and schema for validation or mapping to other systems.Would you like to know more? Click here to find out more details!

FME Hub user vhruska just uploaded a new transformer to the FME Hub.Coda Table WriterThis transformer lets you connect FME to Coda and update table rows using Coda’s Upsert API.👉 Generate your Coda API key hereHow It WorksKey Columns: Choose one or more fields that uniquely identify a row in your Coda table (e.g., ProjectID).Value Columns: Specify which fields should be updated.Upsert Logic: If a row with matching key values exists → it is updated with new values. If no match is found → a new row is created in the table.PerformanceFeatures are processed in batch mode by default, reducing the number of API calls and improving performance for large datasets.Would you like to know more? Click here to find out more details!

FME Hub user crystalwang just uploaded a new transformer to the FME Hub.AzureAIFoundryConnector Transformer DocumentationThe AzureAIFoundryConnector is a custom FME transformer that enables integration with Azure OpenAI's Responses API. This transformer allows you to send prompts to large language models deployed in Azure AI Foundry and receive responses directly within an FME workspace. This transformer also supports a variety of tasks: text, reasoning (only available for some models), image (Vision), and file search.PurposeThis transformer is designed for content generation, structured information extraction, summarization, reasoning, and other AI-driven tasks using Microsoft Azure’s hosted OpenAI services. It supports advanced reasoning models and file-based document processing via Azure AI's Files API.Example Use CaseImagine you have a dataset with a column of customer support inquiries, and you'd like to auto-generate a response or classify the sentiment. You can:Use AttributeCrea

FME Hub user siennaatsafe just uploaded a new transformer to the FME Hub.The MapillaryConnector custom transformer provides access to Mapillary's API to retrieve geospatially referenced imagery and data for use in FME. This transformer enables automation of street-level imagery analysis and extraction of computer vision features directly into your FME workflows.You can use this transformer to:- Retrieve images captured at or near specific geographic locations.- Access detected objects (e.g., traffic signs, poles, crosswalks) via the Object Detection API.- Extract Map Features for spatial analysis and visualization.WARNINGThis tranformer does not support pagination and may not return all results if the record count exceeds 2000. Would you like to know more? Click here to find out more details!

FME Hub user buckel just uploaded a new transformer to the FME Hub.Ce transformeur récupère les données relatives aux entreprises et établissements répertoriés dans le registre administratif français SIRENE, géré par l’INSEE, y compris les entités fermées.Le service est limité à 30 requêtes par minutes.Nécessite FME version 2024.1.0.0 (Build 24612) ou une version plus récente.L’utilisation de ce transformateur nécessite un compte sur le portail de l’INSEE, permettant de créer une application et de s’abonner à l’API SIRENE.Les services actuellement disponibles interrogent :Unités légales (siren)Etablissements (siret)This transformer retrieves data related to businesses and establishments listed in the French administrative SIRENE registry, managed by INSEE, including closed entities.Requires FME version 2024.1.0.0 ( Build 24612 ) or newer.Using this transformer requires an INSEE portal account, which allows you to create an application and subscribe to the SIRENE API.The service is limi

FME Hub user buckel just uploaded a new transformer to the FME Hub.Ce transformeur utilise le service de géocodage inverse de l’IGN pour retourner l’entité ou les entités géolocalisées les plus proches, correspondant à une ou plusieurs des catégories suivantes : adresses, toponymes, parcelles cadastrales et unités administratives.Le service est limité à 50 requêtes par seconde.Nécessite FME version 2024.1.0.0 (Build 24612) ou une version plus récente.This transformer uses the IGN reverse geocoding service to return the nearest geolocated entity or entities matching one or more of the following categories: addresses, place names, cadastral parcels, and/or administrative units.It is limited to 50 requests per second.Requires FME version 2024.1.0.0 ( Build 24612 ) or newer.INPUTLe Transformer à pour but de retourner, à partir d’un ou plusieurs points géographiques indiqués en latitude/longitude, la ou les entités géolocalisées les plus proches correspondantes, parmi les adresses, toponyme

FME Hub user peter_s just uploaded a new web_connection to the FME Hub.This FME Web Service can be used for creating FME Named Connections for the DAF GraphQL EndpointWould you like to know more? Click here to find out more details!

FME Hub user alexbiz just uploaded a new transformer to the FME Hub.The DateTimeFilter custom transformer allows you to filter features based on a date/time attribute.It supports multiple filtering modes to test whether an entity’s date value falls before, after, within, or around specified ranges.FunctionalityThe transformer evaluates the value of a selected DateTime attribute (in FME Date/Time format) against user-defined conditions.Depending on the chosen filter type, features are routed to either:Passed: if the feature’s date meets the filter condition. Failed: if the feature’s date does not meet the filter condition. Filter TypesAfter Date: features with dates strictly after the specified date. Before Date: features with dates strictly before the specified date. Between Dates: features with dates between two specified dates (inclusive). From Date + Duration: features with dates occurring after a given start date, for a specified duration (in minutes). Until Date - Duration:

FME Hub user alexbiz just uploaded a new transformer to the FME Hub.The LogNormalizer custom transformer is designed to read and normalize log files.It attempts to extract and standardize key information such as the timestamp and severity level, producing structured attributes that can be used for further processing in FME.Originally developed to handle FME Flow log files, it can also be applied to log files from other applications.How It WorksEach input feature triggers the parsing of a log file. Typically, it is combined with the Directory and File Pathnames reader to list all log files within a directory and its subdirectories. The transformer uses different regular expressions to detect:The severity level of each log line The timestamp format When a timestamp is found, it is converted and normalized into both FME and ISO 8601 formats. If no timestamp is detected for a line, the transformer attempts to reuse the timestamp of the previous line and marks it as "unknown".Parameter

FME Hub user alexbiz just uploaded a new transformer to the FME Hub.The DateRangeFilter custom transformer filters features based on a timestamp attribute and a user-defined date range.It is designed to work with attributes containing timestamps in the FME Date/Time format.OverviewEach input feature is evaluated against the specified date range.Features with timestamps inside the range are routed to the Inside output port, while those outside are routed to the Outside port.This transformer is useful for workflows where only a subset of features within a certain time window should be processed.ParametersDateTime Attribute (FME Format) Select the attribute that contains the timestamp to evaluate. The timestamp must be in FME Date/Time format.From The start date of the filtering range. Features with timestamps earlier than this date will be routed to Outside.To The end date of the filtering range. Features with timestamps later than this date will be routed to Outside.Output PortsIn

FME Hub user alexbiz just uploaded a new transformer to the FME Hub.The LogParser custom transformer is designed to read and normalize log files.It attempts to extract and standardize key information such as the timestamp and severity level, producing structured attributes that can be used for further processing in FME.Originally developed to handle FME Flow log files, it can also be applied to log files from other applications.How It WorksEach input feature triggers the parsing of a log file. Typically, it is combined with the Directory and File Pathnames reader to list all log files within a directory and its subdirectories. The transformer uses different regular expressions to detect:The severity level of each log line The timestamp format When a timestamp is found, it is converted and normalized into both FME and ISO 8601 formats. If no timestamp is detected for a line, the transformer attempts to reuse the timestamp of the previous line and marks it as "unknown".ParametersLog

FME Hub user andreas_h just uploaded a new transformer to the FME Hub.StaticWebCrawlerThis Custom Transformer is a web crawler designed to extract links from anchor tags and iframes within a specified domain. It respects common guidelines for web crawlers, ensuring ethical crawling by respecting robots.txt rules. The crawler can also discover and utilize sitemap XML files to prioritize crawling based on URL priority values.Key Features:Robots.txt Compliance: Automatically fetches and respects Disallow and Allow rules from a domain's robots.txt file.Sitemap Integration: Discovers sitemaps from robots.txt and uses them to prioritize URLs for crawling based on their <priority> value.Domain Scoping: Restricts crawling to a specific domain to prevent the crawler from wandering.Configurable Limits: Set a maximum number of pages to crawl and a delay between requests to be respectful to servers.Comprehensive Output: Provides detailed attributes for each discovered link, including its sou

FME Hub user alexbiz just uploaded a new transformer to the FME Hub.The UUIDv5Generator is a custom transformer that generates Version 5 Universally Unique Identifiers (UUIDs) for incoming features.UUID v5 values are generated deterministically based on a given namespace and name, making them consistent for the same input parameters. This transformer is useful for creating reproducible identifiers across datasets, workflows, or systems. ParametersNamespaceType: Choice (Dropdown) Options: DNS → Predefined namespace for DNS names URL → Predefined namespace for URLs OID → Predefined namespace for Object Identifiers X500 → Predefined namespace for X.500 Distinguished Names Custom → Use a user-defined namespace Custom NamespaceType: Text (String) Enabled only if: Namespace = Custom Must contain a valid UUID (format: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx). Allows you to define your own namespace for UUID v5 generation. NameType: Text (String) The input string used in combination

FME Hub user andreas_h just uploaded a new transformer to the FME Hub.Origo Config Layer ExtractorSummaryExtracts detailed layer information from an Origo-map configuration JSON file. This transformer is specifically designed to handle Origo's non-standard JSON format, which may include single-line (//) and multi-line (/* ... */) comments, before parsing the layer definitions.GitHubhttps://github.com/origo-mapDetailed DescriptionOrigo map configurations are stored in JSON files that define groups, layers, sources, styles, and more. This transformer simplifies the process of reading these configurations by targeting the layer definitions directly.A key feature of this transformer is its ability to parse JSON files containing comments. Standard JSON parsers will fail on files with // or /* ... */ comments, a common practice in Origo configurations. This transformer pre-processes the file to strip out comments, ensuring robust and error-free parsing.It iterates through all layers defined

FME Hub user virtualcitymatt just uploaded a new transformer to the FME Hub.SafeFilenameCreatorDescriptionThe SafeFilenameCreator transformer ensures that attribute values are converted into safe filenames by replacing all illegal characters with an underscore (_). It works by internally using a StringReplacer to sanitize the attribute values.ParametersAttributes to Replace (String, Required): The name of the attribute(s) whose values will be converted into a safe filename.BehaviorTakes the value of the specified attribute.Replaces all characters that are not allowed in filenames (such as \/:*?"<>|) with _.Outputs the sanitized string as the same attribute value or a new attribute (depending on configuration).Example UsageInput Attribute- Name: Filename- Value: My:Illegal/FileName?.txtOutput Attribute- Name: Filename- Value: My_Illegal_FileName_.txtNotesOnly affects the characters that are illegal in filenames.Can be combined with other transformers to build fully automated file

FME Hub user pierrekoelich just uploaded a new transformer to the FME Hub.IfcGlobalIDGeneratorThis custom transformer creates a field called ifc_global_id and generates a compliant IFC global ID for each feature that is fed in.Would you like to know more? Click here to find out more details!

FME Hub user daveatsafe just uploaded a new transformer to the FME Hub.Encodes point data into the DGGS JSON schema.A DGGS JSON dataset consists of a top zone whose boundary encompasses all the point data and a full set of subzones one or more zoom levels down from the top zone. Each subzone conatins a single set of values for that location. This is conceptually similar to raster data, where the top zone is the raster extents, and the subzones are the pixels. The transformer will group the source points by subzone and calculate the user's choice of mean, median, minimum or maximum to assign to the value of the subzone.How does it work?The DGGSJSONEncoder uses the open source dggal library to generate the top zone and subzone cells. To install the library, please follow the instructions on Installing Python Packages to FME Form to run the following install:fme python -m pip install dggalThis library provides an interface to the following Discrete Global Grids:GNOSIS Global GridIcosahedr

FME Hub user daveatsafe just uploaded a new transformer to the FME Hub.Decodes DGGS JSON encoded data into geometry and attributesHow does it work?The DGGSJSONDecoder uses the open source dggal library to generate the cell geometry. To install the library, please follow the instructions on Installing Python Packages to FME Form to run the following install:fme python -m pip install dggalThis library provides an interface to the following Discrete Global Grids:GNOSIS Global GridIcosahedral Snyder Equal Area Hexagonal 3 (ISEA3H)Icosahedral Snyder Equal Area Rhombic 9 (ISEA9R)Icosahedral Vertex-oriented Great Circle Equal Area Hexagonal 3 (IVEA3H)Icosahedral Vertex-oriented Great Circle Equal Area Rhombic 9 (IVEA9R)Rhombic Triacontahedron Equal Area Hexagonal 3 (RTEA3H)Rhombic Triacontahedron Equal Area Rhombic 9 (RTEA9R)rHEALPix Global GridDiscrete Global Grids partition the earth into a hierarchical set of polygons which can be denoted by unique identifiers, with the size of the polygon

FME Hub user daveatsafe just uploaded a new transformer to the FME Hub.Calculates the relationships between Base and Candidate DGGS grid cell indices. These are often not easily human-discernable from the index values.How does it work?The DGGSRelator provides easy access within an FME workflow to the relationship queries made available by the open source dggal library. To install the library, please follow the instructions on Installing Python Packages to FME Form to run the following install:fme python -m pip install dggalThis library provides an interface to the following Discrete Global Grids:GNOSIS Global GridIcosahedral Snyder Equal Area Hexagonal 3 (ISEA3H)Icosahedral Snyder Equal Area Rhombic 9 (ISEA9R)Icosahedral Vertex-oriented Great Circle Equal Area Hexagonal 3 (IVEA3H)Icosahedral Vertex-oriented Great Circle Equal Area Rhombic 9 (IVEA9R)Rhombic Triacontahedron Equal Area Hexagonal 3 (RTEA3H)Rhombic Triacontahedron Equal Area Rhombic 9 (RTEA9R)rHEALPix Global GridDiscrete Gl