|
| 1 | +# Features |
| 2 | + |
| 3 | +## CSS/XPATH Generator |
| 4 | + |
| 5 | +Enable via `use_llm_for_locator_proposals`=`false` |
| 6 | + |
| 7 | +An internal CSS/XPATH generator will generate a list of unique locator proposals for all relevant elements of a DOM Tree. |
| 8 | +Relevant elements are decided based on the used Keyword. |
| 9 | +For example, for a keyword `Fill Text`, the `input` and `textarea` elements are more relevant. |
| 10 | +For a keyword `Click`, the `button` , `a` and other `clickable` elements are more relevant. |
| 11 | + |
| 12 | +The generated locators are unique and based on attributes like `class`, `id`, `placeholder` , `type` and `name`. |
| 13 | +Also `text` values and next/previous `siblings` as well as `parent` elements are used to support the unique identification. |
| 14 | + |
| 15 | +## AI Supported Self-Healing |
| 16 | + |
| 17 | +Enable via `use_llm_for_locator_proposals`=`true` |
| 18 | + |
| 19 | +DOM Tree will be sent to a LLM, which will identify a list of locator proposals based on the failed locator. |
| 20 | +Locators will depend on the type of LLM. |
| 21 | + |
| 22 | +LLMs can be selected via environment variables: |
| 23 | + |
| 24 | +* `LLM_TEXT_MODEL` (model used for picking final locator from proposal list) |
| 25 | +* `LLM_LOCATOR_MODEL` (model for generating locator proposals from DOM tree) |
| 26 | + |
| 27 | +## Collect additional information for locator proposals |
| 28 | + |
| 29 | +## Locator Database |
| 30 | + |
| 31 | +Enable via |
| 32 | + |
| 33 | +* `collect_locator_info`: Boolean flag to enable or disable the collection of locator information. Default is false. |
| 34 | +* `use_locator_db`: Boolean flag to enable or disable the use of a locator database for selection of fixed locator. Default is false. |
| 35 | +* `locator_db_file`: Specifies the filename for the locator database. Default is "locator_db.json". |
| 36 | + |
| 37 | +During a successful test run, additional details for each working locator can be collected and stored into a TinyDB database by setting `collect_locator_info` to `true`. |
| 38 | + |
| 39 | +The locator database is just a `.json` file and the location and name can be set via `locator_db_file`. |
| 40 | + |
| 41 | +During a test run in which locators shall be healed, the previously collected details in the TinyDB can be used to make better decision for the healed locator. |
| 42 | +That behavior can be enabled by setting `use_locator_db` to `True`. |
| 43 | + |
| 44 | +## Shadow DOM |
| 45 | + |
| 46 | +## DOM Tree reduction |
| 47 | + |
| 48 | +* `fix`: Specifies the mode of operation, set to "realtime" for real-time healing. Default is "realtime". |
| 49 | +* `collect_locator_info`: Boolean flag to enable or disable the collection of locator information. Default is false. |
| 50 | +* `use_locator_db`: Boolean flag to enable or disable the use of a locator database. Default is false. |
| 51 | +* `use_llm_for_locator_proposals`: Boolean flag to enable or disable the use of a language model for generating locator proposals. If true, locator proposals will be identified from DOM Tree via LLM. If set to false, locator proposals are generated via CSS/XPATH generator. Default is false. |
| 52 | +* `heal_assertions`: Boolean flag to enable or disable the healing of assertions. Default is false. (not implemented yet) |
| 53 | +* `locator_db_file`: Specifies the filename for the locator database. Default is "locator_db.json". |
0 commit comments