search

gephi / gephi-plugins

Repository for Gephi Plugins maintained by the team. Each plugin has it's branch.
270 stars 620 forks source link

Bridging Centrality Plugin #200

Closed santosardr closed 5 years ago

santosardr commented 5 years ago

Bridging Centrality (BC) depicts nodes connected to immediate neighbors that, by turn, are connecting clusters of nodes. BC is the result of Betweenness Centrality multiplied by Bridging Node Centrality. BC can be useful to identify links for groups of nodes. For instance, considering a protein network, BC could point up proteins acting as a bridge between complementary molecular functions. This plugin appends the new data columns to the Gephi data laboratory tab: Betweenness, Bridging Node and Bridging Centrality.

eduramiba commented 5 years ago

Hi, thanks for the pull request.

You should actually base your branch on our master branch, and create a pull request into our master-forge branch.

santosardr commented 5 years ago

Sorry, I'm a kind of lost. I do not know where I get lost, so I no sure about how to fix it. Can you help me?

Anderson Rodrigues dos Santos - Professor Adjunto Universidade Federal de Uberlândia – FACOM – sala 1B120 http://lattes.cnpq.br/3752226356973936 http://www.researchgate.net/profile/Anderson_Santos3/

Em ter, 13 de nov de 2018 às 09:09, Eduardo Ramos notifications@github.com escreveu:

Hi, thanks for the pull request.

You should actually base your branch on our master branch, and create a pull request into our master-forge branch.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/gephi/gephi-plugins/pull/200#issuecomment-438227778, or mute the thread https://github.com/notifications/unsubscribe-auth/Aq5aUNncDEjMmXprHLnCW_vCVKyq4bLjks5uuqhXgaJpZM4YbaLf .

eduramiba commented 5 years ago

Sorry, I was confused, I misread the pull request target being our master. Your request is correct.

santosardr commented 5 years ago

Thank you Ramos, By the way, this is one of the products of a master thesis from a student of mine. Can you advise me about a journal that could accept such kind of software tool publication?

Anderson Rodrigues dos Santos - Professor Adjunto Universidade Federal de Uberlândia – FACOM – sala 1B120 http://lattes.cnpq.br/3752226356973936 http://www.researchgate.net/profile/Anderson_Santos3/

Em ter, 13 de nov de 2018 às 09:58, Eduardo Ramos notifications@github.com escreveu:

Sorry, I was confused, I misread the pull request target being our master. Your request is correct.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/gephi/gephi-plugins/pull/200#issuecomment-438241084, or mute the thread https://github.com/notifications/unsubscribe-auth/Aq5aUI66dGA6sNr75206m4q14UVMd22zks5uurPKgaJpZM4YbaLf .

eduramiba commented 5 years ago

Hi, I don't know much about publications, but you can check some existing Gephi citations: https://gephi.org/users/publications/

https://scholar.google.fr/scholar?cites=692893633449340225

santosardr commented 5 years ago

Thank you. I think is better to analyze some biological data proving the value of the plugin. I will do that. Cheers.

Anderson Rodrigues dos Santos - Professor Adjunto Universidade Federal de Uberlândia – FACOM – sala 1B120 http://lattes.cnpq.br/3752226356973936 http://www.researchgate.net/profile/Anderson_Santos3/

Em ter, 13 de nov de 2018 às 10:07, Eduardo Ramos notifications@github.com escreveu:

Hi, I don't know much about publications, but you can check some existing Gephi citations: https://gephi.org/users/publications/

https://scholar.google.fr/scholar?cites=692893633449340225

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/gephi/gephi-plugins/pull/200#issuecomment-438243569, or mute the thread https://github.com/notifications/unsubscribe-auth/Aq5aUPfD18npcJJ3w_aZwp3UEKSIglJBks5uurYFgaJpZM4YbaLf .

santosardr commented 5 years ago

Dear Eduardo, I built this plugin to manage graphs with more than four thousand nodes and many hundreds of thousand edges, and It's working, Considering this magnitude order, I defined my variables as double since probably the relevant information will be at least ten digits at right, I have not a single variable defined as float. Unfortunately, even calling the method as ... nodeTable.addColumn(BRIDGING_CENTRALITY, "Bridging Centrality", Double.class, (double) 0); Even so, I just can read a precision of at most six digits in the Data Laboratory tab, and the majority of the data are listing just zeros (0.000000), and a few lines shows something different of zeros. So, If I didn't anything wrong I believe this is a limitation of your primary classes. Please, Can you help me?

Regards, Anderson Santos

Anderson Rodrigues dos Santos - Professor Adjunto Universidade Federal de Uberlândia – FACOM – sala 1B120 http://lattes.cnpq.br/3752226356973936 http://www.researchgate.net/profile/Anderson_Santos3/

Em ter, 13 de nov de 2018 às 09:58, Eduardo Ramos notifications@github.com escreveu:

Sorry, I was confused, I misread the pull request target being our master. Your request is correct.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/gephi/gephi-plugins/pull/200#issuecomment-438241084, or mute the thread https://github.com/notifications/unsubscribe-auth/Aq5aUI66dGA6sNr75206m4q14UVMd22zks5uurPKgaJpZM4YbaLf .

eduramiba commented 5 years ago

Mmm I think data laboratory limits precision, can you try with BigDecimal.class? Anyway, on double click it should show the full value.

santosardr commented 5 years ago

I did not know such tip (double-click). Indeed, the full precision is shown with this action, so I don't think in varying the number will change de data laboratory precision. The worst happens when we export the data, the number 0.00011 is showing as 1.1034059555675141E-4 after the double-click, but after export to CSV, it is just 0.00011. The point here is the user must double-click a lot of cells to have a clue about the data distribution, for instance, to select the relevant top 20 nodes that are not exactly zero. Please, fix this.

Anderson Rodrigues dos Santos - Professor Adjunto Universidade Federal de Uberlândia – FACOM – sala 1B120 http://lattes.cnpq.br/3752226356973936 http://www.researchgate.net/profile/Anderson_Santos3/

Em qua, 14 de nov de 2018 às 17:10, Eduardo Ramos notifications@github.com escreveu:

Mmm I think data laboratory limits precision, can you try with BigDecimal.class? Anyway, on double click it should show the full value.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/gephi/gephi-plugins/pull/200#issuecomment-438781062, or mute the thread https://github.com/notifications/unsubscribe-auth/Aq5aUOPbeLaWGqDWx5Tp2sSl3HmLh5ulks5uvGqbgaJpZM4YbaLf .

eduramiba commented 5 years ago

I guess we should increase the shown precision to 5 or 6 decimals, and show full precision with bigdecimal, but that will need to wait for a next release.

Anyway, I advise you to use BigDecimal if you want some real precision, as floats and doubles have inherent precision loss, that's why we round it.

santosardr commented 5 years ago

Thank you for your answer, Eduardo. I will make the change for BigDecimal and will be looking forward to the next release. BTW, If such release will happen before 2019, March I will be pleased because this is my planning date to release a publication citing GEPHI as the tool used in the methods.

Cheers, Anderson Santos

Anderson Rodrigues dos Santos - Professor Adjunto Universidade Federal de Uberlândia – FACOM – sala 1B120 http://lattes.cnpq.br/3752226356973936 http://www.researchgate.net/profile/Anderson_Santos3/

Em qua, 14 de nov de 2018 às 17:31, Eduardo Ramos notifications@github.com escreveu:

I guess we should increase the shown precision to 5 or 6 decimals, and show full precision with bigdecimal, but that will need to wait for a next release.

Anyway, I advise you to use BigDecimal if you want some real precision, as floats and doubles have inherent precision loss, that's why we round it.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/gephi/gephi-plugins/pull/200#issuecomment-438788991, or mute the thread https://github.com/notifications/unsubscribe-auth/Aq5aUOz6u2LzmGIDVyi_3U05wucIj70dks5uvG-igaJpZM4YbaLf .

santosardr commented 5 years ago

Eduardo, I thought you were guessing about the BigDecimal. Sorry. Yes, it did the trick. [image: Captura de tela de 2018-11-14 20-43-39.png] The selected node is the last one. The above-listed data is a small sample with just one hundred nodes, and I already have a lot of zeros. I will push this modification. However, the CSV export still has the problem. BTW, I need to make a new pull request for you be able to incorporate the changes?

Anderson Rodrigues dos Santos - Professor Adjunto Universidade Federal de Uberlândia – FACOM – sala 1B120 http://lattes.cnpq.br/3752226356973936 http://www.researchgate.net/profile/Anderson_Santos3/

Em qua, 14 de nov de 2018 às 17:37, Anderson Santos santosardr@gmail.com escreveu:

Thank you for your answer, Eduardo. I will make the change for BigDecimal and will be looking forward to the next release. BTW, If such release will happen before 2019, March I will be pleased because this is my planning date to release a publication citing GEPHI as the tool used in the methods.

Cheers, Anderson Santos

Anderson Rodrigues dos Santos - Professor Adjunto Universidade Federal de Uberlândia – FACOM – sala 1B120 http://lattes.cnpq.br/3752226356973936 http://www.researchgate.net/profile/Anderson_Santos3/

Em qua, 14 de nov de 2018 às 17:31, Eduardo Ramos < notifications@github.com> escreveu:

I guess we should increase the shown precision to 5 or 6 decimals, and show full precision with bigdecimal, but that will need to wait for a next release.

Anyway, I advise you to use BigDecimal if you want some real precision, as floats and doubles have inherent precision loss, that's why we round it.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/gephi/gephi-plugins/pull/200#issuecomment-438788991, or mute the thread https://github.com/notifications/unsubscribe-auth/Aq5aUOz6u2LzmGIDVyi_3U05wucIj70dks5uvG-igaJpZM4YbaLf .

santosardr commented 5 years ago

Erratum: the sent image does not show the last node but believe me, it's very small.

Anderson Rodrigues dos Santos - Professor Adjunto Universidade Federal de Uberlândia – FACOM – sala 1B120 http://lattes.cnpq.br/3752226356973936 http://www.researchgate.net/profile/Anderson_Santos3/

Em qua, 14 de nov de 2018 às 20:45, Anderson Santos santosardr@gmail.com escreveu:

Eduardo, I thought you were guessing about the BigDecimal. Sorry. Yes, it did the trick. [image: Captura de tela de 2018-11-14 20-43-39.png] The selected node is the last one. The above-listed data is a small sample with just one hundred nodes, and I already have a lot of zeros. I will push this modification. However, the CSV export still has the problem. BTW, I need to make a new pull request for you be able to incorporate the changes?

Anderson Rodrigues dos Santos - Professor Adjunto Universidade Federal de Uberlândia – FACOM – sala 1B120 http://lattes.cnpq.br/3752226356973936 http://www.researchgate.net/profile/Anderson_Santos3/

Em qua, 14 de nov de 2018 às 17:37, Anderson Santos santosardr@gmail.com escreveu:

Thank you for your answer, Eduardo. I will make the change for BigDecimal and will be looking forward to the next release. BTW, If such release will happen before 2019, March I will be pleased because this is my planning date to release a publication citing GEPHI as the tool used in the methods.

Cheers, Anderson Santos

Anderson Rodrigues dos Santos - Professor Adjunto Universidade Federal de Uberlândia – FACOM – sala 1B120 http://lattes.cnpq.br/3752226356973936 http://www.researchgate.net/profile/Anderson_Santos3/

Em qua, 14 de nov de 2018 às 17:31, Eduardo Ramos < notifications@github.com> escreveu:

I guess we should increase the shown precision to 5 or 6 decimals, and show full precision with bigdecimal, but that will need to wait for a next release.

Anyway, I advise you to use BigDecimal if you want some real precision, as floats and doubles have inherent precision loss, that's why we round it.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/gephi/gephi-plugins/pull/200#issuecomment-438788991, or mute the thread https://github.com/notifications/unsubscribe-auth/Aq5aUOz6u2LzmGIDVyi_3U05wucIj70dks5uvG-igaJpZM4YbaLf .

eduramiba commented 5 years ago

Hi Anderson, Finally your plugin is merged and should be available soon. Thank you very much for the effort and the patience for this to be merged.

I made some small corrections to your code, so please be sure to merge the following pull request into your codebase so you can later update the plugin: santosardr#1

eduramiba commented 5 years ago

Hi, Just noticed that you changed the root pom.xml in your branch and I merged it by accident. To clarify, plugins request should never touch other poms that their own module poms.

eduramiba commented 5 years ago

Your plugin is now available at https://gephi.org/plugins and in Gephi - Tools - Plugins menu

santosardr commented 5 years ago

Thank you for the excellent news Eduardo.

Best regards, Andeson Santos

Anderson Rodrigues dos Santos - Professor Adjunto Universidade Federal de Uberlândia – FACOM – sala 1B120 http://lattes.cnpq.br/3752226356973936 http://www.researchgate.net/profile/Anderson_Santos3/

Em sáb, 5 de jan de 2019 às 11:34, Eduardo Ramos notifications@github.com escreveu:

Your plugin is now available at https://gephi.org/plugins and in Gephi - Tools - Plugins menu

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/gephi/gephi-plugins/pull/200#issuecomment-451655715, or mute the thread https://github.com/notifications/unsubscribe-auth/Aq5aUKSrnEvdgbMzo96qCVlLViILcaJMks5vAKnggaJpZM4YbaLf .