The plugins library for jerlendds/osintbuddy.
output.webm
Please note: OSINTBuddy plugins are still in early alpha and breaking changes may occasionally occur in the API. That said, if you remain on the main
branch and avoid accessing protected methods we will try our best to avoid introducing breaking changes.
NOTICE: There has been a major update to plugins, any created plugins will have to be updated to use the new, and more convenient data access method:
- Remove
name
from anyob.Plugin
- Update
node['data']
access to benode.label_defined_in_node
- Please see the introduction to the plugin system below
The osintbuddy plugin system at its core is very simple. An OBRegistry
class holds all registered OBPlugin
classes within the application. This registry is loaded into the osintbuddy application where it is then used to load the available entities for the user when they access a project graph, load transforms when a user opens the context menu of a node, and perform transformations which expect a Plugin.blueprint()
to be returned. The returned data of a transform decorated method will be automatically mapped to a JanusGraph database through gremlin according to the labels (as snakecase) you previously set in the classes node
for whatever Plugin.blueprint()
you return.
To make this a bit more clear please see the below example...
from pydantic import BaseModel
import osintbuddy import transform, Plugin
from osintbuddy.elements import TextInput, DropdownInput, Title, CopyText
from osintbuddy.errors import OBPluginError
class CSESearchResults(Plugin):
label = "CSE Result"
name = "CSE result"
show_label = False # do not show this on the entities dialog
# the user sees on the left of the project graph screen
color = "#058F63"
node = [
Title(label="Result"),
CopyText(label="URL"),
CopyText(label="Cache URL"),
]
class CSESearchPlugin(Plugin):
label = "CSE Search"
name = "CSE search"
color = "#2C7237"
node = [
[
TextInput(label="Query", icon="search"),
TextInput(label="Pages", icon="123", default="1"),
],
DropdownInput(label="CSE Categories", options=cse_link_options)
]
@transform(label="To cse results", icon="search")
async def transform_to_cse_results(
self,
node: BaseModel, # dynamically generated pydantic model
# that is mapped from the above labels contained within `node`
use # a pydantic model allowing you to access a selenium instance
# (and eventually a gremlin graph and settings api)
):
results = []
if not node.query:
raise OBPluginError((
'You can send error messages to the user here'
'if they forget to submit data or if some other error occurs'
))
# notice how you can access data returned from the context menu
# of this node; using the label name in snake case
print(node.cse_categories, node.query, node.pages)
... # (removed code for clarity)
if resp:
for result in resp["results"]:
url = result.get("breadcrumbUrl", {})
# some elements you can store more than just a string,
# (these elements storing dicts are mapped
# to janusgraph as properties with the names
# result_title, result_subtitle, and result_text)
blueprint = CSESearchResults.blueprint(
result={
"title": result.get("titleNoFormatting"),
"subtitle": url.get("host") + url.get("crumbs"),
"text": result.get("contentNoFormatting"),
},
url=result.get("unescapedUrl"),
cache_url=result.get("cacheUrl"),
)
results.append(blueprint)
# here we return a list of blueprints (blueprints are dicts)
# but you can also return a single blueprint without a list
return results