VisJS network interface#

VisJS offers nice graph displays.

gismap.gisgraphs.builder.make_vis(lab, **kwargs)[source]#

Generate HTML visualization of a lab’s collaboration network.

Parameters:

lab (LabMap) – Lab to display.
uid (str, optional) – Unique identifier for DOM elements. Auto-generated if not provided.
vis_url (str, optional) – URL to vis-network library.
groups (dict, optional) – Group styling configuration.
draw_legend (bool, optional) – Whether to draw the legend. Defaults to True if multiple groups.
physics (dict, optional) – Physics engine configuration.
nodes_options (dict, optional) – Node styling options.
edges_options (dict, optional) – Edge styling options.
interaction_options (dict, optional) – Interaction settings.
style (string.Template, optional) – CSS template.
script (string.Template, optional) – JavaScript template.
theme (str, default=”auto”) – Initial color theme: "auto" follows the host (Jupyter / Sphinx), "light" or "dark" force a palette. Switchable at runtime from the menu.
inline_images (bool, default=True) – If True, fetch each node image and embed it as a data: URI so the canvas stays clean and PNG export works. Images that fail to download or exceed max_inline_bytes are demoted to the initials fallback (no external URL is ever left on a node).
max_inline_bytes (int, default=200_000) – Skip inlining for any image larger than this (e.g. uncropped LDAP portraits). The node is demoted to initials.

Returns:

HTML code as a string.

Return type:

str

gismap.gisgraphs.graph.initials(name)[source]#

Parameters:: name (str) – Person’s name.
Returns:: Person’s initials. Composite first names keep all their initials, while the common First Last case stays two letters.
Return type:: str

Examples

>>> initials("Fabien Mathieu")
'FM'
>>> initials("Jean-François Laslier")
'JFL'
>>> initials("Jérôme Lang")
'JL'
>>> initials("Élie de Panafieu")
'ÉP'
>>> initials("Madonna")
'M'

gismap.gisgraphs.graph.lab_to_graph(lab)[source]#

Parameters:: lab (LabMap) – A lab populated with authors and publications.
Returns:: (nodes, edges, publications) where nodes and edges carry only display data plus pub_keys references, and publications is a dict keyed by publication key with the full per-pub payload (title, authors, venue, year, url, abstract, bib). Modal HTML is built JS-side from this shared dict, so each publication is shipped only once even when it touches many authors and pairs.
Return type:: tuple

Examples

>>> from gismap.lab import ListMap as Map
>>> lab = Map(author_list=['Tixeuil Sébastien', 'Mathieu Fabien'], name='mini', dbs="hal")
>>> lab.update_authors()
>>> lab.update_publis()
>>> len(lab.authors)
2
>>> 320 < len(lab.publications) < 430
True
>>> nodes, edges, pubs = lab_to_graph(lab)
>>> nodes[0]['group']
'mini'
>>> edges[0]['hover']
'Show joint publications from Mathieu Fabien and Tixeuil Sébastien'
>>> sample = next(iter(pubs.values()))
>>> {'title', 'authors', 'bib'} <= set(sample)
True
>>> html = lab.html(groups={"mini": {"color": "#777"}})

gismap.gisgraphs.graph.node_labels(names, max_len=5)[source]#

Compute display labels (initials) for a set of authors, disambiguating collisions.

Parameters:

names (dict) – Mapping key -> full name.
max_len (int, default=5) – Hard cap on label length. Disambiguation stops extending a label once it reaches this length, even if a collision remains — this keeps labels readable when names cannot be told apart (e.g. a source that yields “Surname Firstname” makes shared first names look like a shared surname). Set higher to allow longer disambiguation suffixes.

Returns:

Mapping key -> label. When several authors share the same base initials, their labels are extended with lowercase letters from the surname until distinct (or until max_len is reached).

Return type:

dict

Examples

>>> labels = node_labels({1: "Jérôme Lang", 2: "Julien Lesca", 3: "Jean-François Laslier"})
>>> [labels[k] for k in (1, 2, 3)]
['JLa', 'JLe', 'JFL']
>>> node_labels({1: "Fabien Mathieu"})[1]
'FM'

gismap.gisgraphs.graph.to_edge(k, pub_keys, searchers)[source]#

Parameters:

k (tuple) – Keys of the two authors.
pub_keys (list) – Joint publication keys, year-desc order.
searchers (dict) – Authors keyed by author key.

Returns:

Display-ready data for the collaboration edge.

Return type:

dict

gismap.gisgraphs.graph.to_node(s, pub_keys, label=None)[source]#

Parameters:

s (LabAuthor) – Author.
pub_keys (list) – Publication keys associated with this author, year-desc order.
label (str, optional) – Pre-computed (and possibly disambiguated) initials for this node. Falls back to initials() when not provided.

Returns:

Display-ready data for the node. Modal HTML is rendered on click by the JS layer using pub_keys and the shared publications dict.

Return type:

dict

gismap.gisgraphs.groups.is_ego(lab)[source]#: True when lab is an EgoMap (checked by name to avoid a circular import).

class gismap.gisgraphs.widget.GismapWidget[source]#

A simple widget to test the production of LabMaps and EgoMaps.

Examples

This is a doctest example. Use a notebook to play with the widget.

>>> gw = GismapWidget()
VBox(children=(HTML(value=''), Output(), HBox(children=(Textarea(value='', ...
>>> gw.names.value = "Fabien Mathieu"
>>> gw.dbs.value = "HAL"
>>> gw.size.value = 3
>>> gw.compute_function(gw.compute, show=False)
>>> gw.save_link.value[:30]
"<a href='data:text/html;base64"
>>> gw.names.value = "Diego Perino, Laurent Viennot"
>>> gw.compute_function(gw.compute, show=False)
>>> gw.save_link.value[:30]
"<a href='data:text/html;base64"

compute_function(b, show=True)[source]#

Handle compute button click and generate visualization.

Parameters:

b (ipywidgets.Button) – The button widget that triggered this callback.
show (bool, default=True) – Whether to display the result in the widget.

Return type:

None

html()[source]#

Generate the HTML visualization based on widget inputs.

Returns:: HTML content for the collaboration graph.
Return type:: str

gismap.gisgraphs.widget.safe_filename(name)[source]#

Parameters:: name (str) – Pretty much anything.
Returns:: GisMap filename.
Return type:: str