1 GENERAL CONCEPTS of NLG and the platform (Intro)
//Präambel, What's the platform for? How is it technically structured? What are its components? What are the basic functions? To discuss/ ATML3, Keywords (SEO), Ruleset Authoring//
2 GENERAL SETTINGS (Profile & Team Management)
2.1 Settings - Manage Team:
//What is a team on the platform? What do roles mean (text in UI) How can I change roles (as an owner)? What else can I do here? Switch between teams//
2.2 Settings - API Access
Link to API Documentation (Text in UI)
2.3 ADVANCED SETTINGS
(explained in the UI: Activity Log, Lexicon, Histogram, axite, project-wide message list)
3. UPLOAD & REPORTS
//What can I do/see here?//
4.1 Project Settings - Licenses
4.2 Project Settings - Delete Projects
//What does delete a project mean?//
4.3 Instant Generation Endpoints (text in UI): Link to API Docs
5 DATA SOURCE/COLLECTION
//technical criteria for data quality (?) Create (explain: Language setting, what does this mean?) Edit Collection: Settings, delete/empty collection Preview Test Object//
5.1 IMPORT DATA
//Intro: Preview Language (Translation), Configure Language, Preview Test Object
Analyze now Filters (field filling, unique fields Field Data Types Field values Add data nodes (?)
What to do in the Transform, the functionality: aggregation, calculating, filter, lambda, vertigo (explain)
Description, function, Colorcode Ports Connections
9.1 Value Access Methods
//import document Value Access Methods//
9.2 Special type: VARIABLES
Variables differ from the other nodes in their graphical representation and functionality: They are the only nodes that can be used directly in the text. They therefore always form the end point of a sequence. And any other node (sequence) needs this variable endpoint, if it is to use in the text. For using in the text, Variable nodes are also directly available in the Write Tab.
Variable nodes as triggers Variable nodes are also used as triggers for statements, branches and stories. For this they are connected to a condition node, where the the condition under which an element should appear is formulated.
//explain for every type: applications, functions, parameters,//
9.3 DATA Nodes
Generate Variable Definition/explain Data nodes Define the Types: String, Number, Bolean, List, Split list, frequent Values Delimiter DATA NODE Data Type
9.4 MAPPING Mapping (Code)
what is mapping? mapping expressions (list?)
9.5 CONDITION NODE
Language Noun Adjective Headnoun
Lookup table Default value key
9.10 Structural and Logical Operators for Nodes
// import documents Structural Operations and logical Operators//
10 EXPRESSIONS (oder 9.11?)
// import documentation "Expressions"
11 MANAGEMENT TOOLS in Transform
Subgraphs are an embedded work sheet where you can place your nodes in our Transform section. They allow you to keep your configurations clean and group certain nodes together. You can create large structures of data handling inside a subgraph and use its results outside without cluttering your main graph. Subgraphs can also be moved, copied and pasted as often as you like and where you like (also to other projects!) How to add ports to subgraphs
- Add a subgraph in the Transform section of the COMPOSER
- In the right menu, find Input Connectors and Output Connectors
- Click on ADD CONNECTOR in either section to add Ports to the subgraph. Error (Error message) LookupTable Keys add phrases Edit heads (Noun, Adjective, Headnoun)
12 WRITE Intro
//What to do in the WRITE, the functionality//
Writing with example data set, instant preview, Preview Test object Add Trigger (light bulb symbol) CONTAINER Definition and functions of Container What to mark as container /how to annotate Automatic annotation
//Concept, Annotate, Functions//
14.1 CONTAINER SETTINGS (Alex D)
Variable Output Variable Output Content: Phrase Parts (determiner), Use Phrase Adjectives, Use Numeral , Output whole phrase as a pronoun Variable Output Grammar: contextual Grammar, Lexical Grammar Variable Output Trigger: Trigger Variable, Conditions, Alternative Content Variable Output Formatting: static parts, Output Text Formatting ROLES GRAMMAR Grammar Provider Grammar Provider Noun/Verb: Lemma, Phrase Parts (s.o.)
Static Text Static Text Synonyms
//branching and branches branching settings: Trigger, weight skip branch if variables are empty//
Branches are visual representations of the structures of your statements. Branches give support at formulating your statements in more variable ways and arranging their components according your requirements. When writing a statement you have the possibility to branch off at certain points in a statement to create more than one way to express the information or to activate different predicatements, depending on the data. For this you can create branches that span one or more words, phrases, or entire sentences. The branches appear one below the other in the Write Tab, so you can easily keep track of your construction . For each branch you can specify in the Branching Point different modes of deciding which branch to render, in this way you can easily implement logic while writing your text. Within a branch (sub)-branches can be created. You can copy and paste branches for multiple use.
Functions of branches:
- Write multiple variants of statements and see them all at once .
- Set and manage synonyms – words, sentence parts or complete sentences for variation in your texts.
- Define a proper verbal output for different data .
- Set conditions for rendering certain parts of statements.
- Adjust the variance of the branches very flexibly by setting a weight parameter.
- Refer to container by role.
The settings for branches and branching You create branches in the Write Tab of the NLG Platform by marking these parts of a statement, that should have more than one variant (and clicking on the Branching Symbol) With this marking you define a point where the statements will split into one ore more branches, called branching and the point where the branches come together again.
For each branching you can define the conditions under which this branch should appear
- Select the corresponding trigger variable you created in the Transform Tab .
- weight Depending on the specification of the weight, the probability that the branch is triggered increases or decreases.
When you click on the Branching Symbol you open the settings of the branching in the right panel, where you can define the mode of processing the associated branches:
- first: the first valid branch will be chosen
- random: the branches are chosen randomly ()
16 LEXICON (Text in UI)
missing lexicon entries global lexicon lemma, language, appearance, date, part of speech verbs and nouns switch determiner switch preposition
17 LOOKUPS (see also in Nodes)
18 NARRATE: STORIES
What is a story? Building blocks for stories: Statements Triggers for stories trigger Statements (always, trigger, obligatory or not) Delete statements/stories (
What function has the Review Tab (What happens here? What to do?)
Collection Export Generate Filtered error nodes Generate all
Properties (UID, UUID, Training hash, text order, missing lexicon entries) Imported data Metrics ATML-Debug