Rather a request for advice than an issue… about fcapy HOT 6 CLOSED

Hello Yannis!
I have not run your code yet. But I think the misconception is caused simply by the way I denote the nodes.
(And, I think, this way is default in FCA community. See https://upriss.github.io/fca/examples.html ).

Given a concept, its intent can be derived by union of attributes that are shown in concepts above the current concept. And the extent of the concept can be derived by union of objects that are shown in concepts below the current concept.
For example, for concept # 3, the intent would be {eggs, dangerous} (where "eggs" is shown in node 3 and "dangerous" is shown in node 1). The extent would be {alligator, tarantula} (where "alligator" is shown in node 6 and "tarantula" is shown in node 8).
You should get the same extent and intent of concept 3 by writing:

print(L[3].extent)
print(L[3].intent)

Visualizing the whole extent and the whole intent for each node would just overload the diagram with lots of text.

Now, if you do not like such notation, you can define you own via node_label_func parameter of draw_concept_lattice function. node_label_func should be a python function that takes an index of the concept el_i, the lattice ltc, and outputs a text to visualize.
For example:

vsl.draw_concept_lattice(
    L, ax=ax, flg_node_indices=True,
    max_new_extent_count=10, flg_new_intent_count_prefix=False, flg_new_extent_count_prefix=False,
   node_label_func = lambda el_i, ltc:  ','.join(ltc[el_i].intent) + '\n\n\n' + ','.join(ltc[el_i].extent) 
)

This code should draw the lattice with full intent and full extent for each node.

from fcapy.

Please, notify me if the aforementioned code does not work. Then I will have to modify the package, which might take some time.

from fcapy.

It does work and it is great, THANK YOU so much. Just one last query: how can I get the empty set symbol for the top and bottom nodes?

from fcapy.

Here, the visualization is done via networkx package which supports latex. So, I think, the following string should work fine

r"$\emptyset$"

Applied to the code above, it would be somewhat like

def full_labels_func(el_i, ltc):
   intent_str = ','.join(ltc[el_i].intent) if len(ltc[el_i].intent) > 0 else r'$\emptyset$'
   extent_str = ','.join(ltc[el_i].extent) if len(ltc[el_i].extent) > 0 else r'$\emptyset$'
   return intent_str + '\n\n\n' + extent_str

vsl.draw_concept_lattice(
    L, ax=ax, flg_node_indices=True,
    max_new_extent_count=10, flg_new_intent_count_prefix=False, flg_new_extent_count_prefix=False,
   node_label_func = full_labels_func
)

from fcapy.

Thank you Egor, you are really very very kind, fast and efficient! I'm very grateful! This diagram will appear in my book on NLP and you will be duly thanked in it!
Yannis

from fcapy.

Thank you! It it such an honour :)

However, if this diagram is going to be published, I would propose to slightly enhance it. For example, move the node 8 somewhere above the node 10, as now it is placed too close to the line between node 7 and node 10.
I have not yet implemented the algorithm for beutiful placing of nodes (and I do not really know one). So I do all prettyfications manually. The example is presented in this tutorial: https://github.com/EgorDudyrev/FCApy_tutorials/blob/main/Lattice%20Visualization/Visualizing_lattice.ipynb
This work will take a bit more time and effort, but the result would look much nicer.

from fcapy.