Reading a page more closely¶
playwright_extract ¶
Take pieces out of a page instead of its whole text: a CSS selector or an XPath in, exactly one, and the text of each element it matches out — or one attribute of each, such as href. For the rows of a table or the links of a list. A page read within 10 minutes is not visited again.
read-only · every call adds · may reach outside services.
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
url |
string |
yes | — | Absolute http or https address of a public page. |
xpath |
string? |
— | null |
XPath 1.0 over the rendered markup, lower-case names: //table[@id='prices']//tr, //h2. |
css |
string? |
— | null |
CSS selector as a browser takes it: table#prices tr, a[href*='/docs/']. |
attribute |
string? |
— | null |
Attribute to return instead of the text, such as href; a relative address comes back absolute. |
limit |
integer |
— | 50 |
Pieces to return at most, 1 to 200; 50 by default. |
JSON Schema
{
"type": "object",
"properties": {
"url": {
"description": "Absolute http or https address of a public page.",
"type": "string"
},
"xpath": {
"description": "XPath 1.0 over the rendered markup, lower-case names: //table[@id=\u0027prices\u0027]//tr, //h2.",
"type": [
"string",
"null"
],
"default": null
},
"css": {
"description": "CSS selector as a browser takes it: table#prices tr, a[href*=\u0027/docs/\u0027].",
"type": [
"string",
"null"
],
"default": null
},
"attribute": {
"description": "Attribute to return instead of the text, such as href; a relative address comes back absolute.",
"type": [
"string",
"null"
],
"default": null
},
"limit": {
"description": "Pieces to return at most, 1 to 200; 50 by default.",
"type": "integer",
"default": 50
}
},
"required": [
"url"
]
}
playwright_extract_fields ¶
Take a record of named values out of a page. Each field tries its selectors in order, css before xpath, and the first that matches gives the value. The answer names the selector that matched each field — keep it and pass it first next time — and lists under missing the fields no selector found: read the page with playwright_fetch and find those anew.
read-only · every call adds · may reach outside services.
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
url |
string |
yes | — | Absolute http or https address of a public page. |
fields |
array |
yes | — | 1 to 30 fields, each { name, css: [...], xpath: [...], attribute, many } with 1 to 5 selectors a field; attribute such as href instead of the text; many for every match. |
JSON Schema
{
"type": "object",
"properties": {
"url": {
"description": "Absolute http or https address of a public page.",
"type": "string"
},
"fields": {
"description": "1 to 30 fields, each { name, css: [...], xpath: [...], attribute, many } with 1 to 5 selectors a field; attribute such as href instead of the text; many for every match.",
"type": "array",
"items": {
"type": "object",
"properties": {
"name": {
"type": "string"
},
"css": {
"type": [
"array",
"null"
],
"items": {
"type": "string"
},
"default": null
},
"xPath": {
"type": [
"array",
"null"
],
"items": {
"type": "string"
},
"default": null
},
"attribute": {
"type": [
"string",
"null"
],
"default": null
},
"many": {
"type": "boolean",
"default": false
},
"candidates": {
"type": [
"array",
"null"
],
"items": {
"type": "object",
"properties": {
"text": {
"type": "string"
},
"kind": {
"type": "string",
"enum": [
"Css",
"XPath"
]
}
},
"required": [
"text",
"kind"
]
}
}
},
"required": [
"name"
]
}
}
},
"required": [
"url",
"fields"
]
}
playwright_screenshot ¶
Look at a page instead of reading it: an address in, a picture of what the browser drew out. For what text cannot carry — a chart, a map, a layout, a page whose content is an image. It is slower and heavier than playwright_fetch and says nothing a picture does not show, so read first and look only when the text is not enough.
read-only · every call adds · may reach outside services.
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
url |
string |
yes | — | Absolute http or https address of a public page. |
wholePage |
boolean |
— | false |
True to picture the page from top to bottom, up to 8000 pixels; false, the default, pictures the screen a visitor first sees. |
JSON Schema
{
"type": "object",
"properties": {
"url": {
"description": "Absolute http or https address of a public page.",
"type": "string"
},
"wholePage": {
"description": "True to picture the page from top to bottom, up to 8000 pixels; false, the default, pictures the screen a visitor first sees.",
"type": "boolean",
"default": false
}
},
"required": [
"url"
]
}
playwright_perform ¶
Use a page instead of only reading it: open an address, make a few moves on it, and get the text of the page that results. For what a page says only after a form is filled in — a registry search, a filter, a site's own search box. A move the site did not react to is reported with what the page offered instead: rewrite the move from that rather than repeat it. It does not sign in and does not solve a challenge.
may overwrite or delete · every call adds · may reach outside services.
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
url |
string |
yes | — | Absolute http or https address of the public page to start on. |
moves |
array |
yes | — | 1 to 12 moves, made in order. action: type (key by key into a field), pick (press the offered suggestion whose text matches value), choose (an option of a real select list, by its visible text), click, or press (a key while the element has the focus; Enter unless value names another). Name the element one way: css, a selector, or label, the visible caption of a field. expect with pattern holds a move to the site's reaction: request (the page asked for a URL containing pattern), appears or vanishes (an element matching the selector in pattern), text (the page came to contain the phrase). withinSeconds is how long that reaction is waited for, 1 to 30; 5 by default. |
maxCharacters |
integer |
— | 20000 |
Characters of the resulting page's text to return, 500 to 100000; 20000 by default. |
JSON Schema
{
"type": "object",
"properties": {
"url": {
"description": "Absolute http or https address of the public page to start on.",
"type": "string"
},
"moves": {
"description": "1 to 12 moves, made in order. action: type (key by key into a field), pick (press the offered\nsuggestion whose text matches value), choose (an option of a real select list, by its visible text),\nclick, or press (a key while the element has the focus; Enter unless value names another). Name the\nelement one way: css, a selector, or label, the visible caption of a field. expect with pattern\nholds a move to the site\u0027s reaction: request (the page asked for a URL containing pattern), appears\nor vanishes (an element matching the selector in pattern), text (the page came to contain the\nphrase). withinSeconds is how long that reaction is waited for, 1 to 30; 5 by default.",
"type": "array",
"items": {
"type": "object",
"properties": {
"action": {
"type": "string"
},
"css": {
"type": [
"string",
"null"
],
"default": null
},
"label": {
"type": [
"string",
"null"
],
"default": null
},
"value": {
"type": [
"string",
"null"
],
"default": null
},
"expect": {
"type": [
"string",
"null"
],
"default": null
},
"pattern": {
"type": [
"string",
"null"
],
"default": null
},
"withinSeconds": {
"type": [
"integer",
"null"
],
"default": null
}
},
"required": [
"action"
]
}
},
"maxCharacters": {
"description": "Characters of the resulting page\u0027s text to return, 500 to 100000; 20000 by default.",
"type": "integer",
"default": 20000
}
},
"required": [
"url",
"moves"
]
}