L'URL pour envoyer la table user est fausse et renvoie une erreur 404. Le problème vient du fait que la chaîne de caractère "users" ne prend en réalité pas de "s" à la fin. La ligne qui pose problème dans la fonction est juste ci-dessous.

#path <- httr::modify_url(server, path = paste0(mangal.env$base, "/users/?name=", users[[1]]))

Transform the font as path with Inkscape.

POST_environment() can only accept one attribute and one value. Could be useful to have a function for posting an environmental dataset like : each row is a different site and each column is an environmental variable.

reprex: get_network_by_id(19)

I've been waiting for >15min, still not done, all data are in, it's a list of >15,000 elements, not that big.

The culprit is the do.call(rbind,) approach for large list of sf objects see r-spatial/sf#798. I'm working on a workaround, basically I'll delay the creation of sf object as much as possible.

Coerce mgNetwork class to igraph object

So far limit is never passed on to the API, so changing it increases the number of pages but those pages do not exist!

• Create mgNetwork class (one network, all interactions at a specific location and time)
• Create mgNetworkCollection (several networks of a dataset)

• Set previous Gab version on a branch (in order to not break mangal-wg/mangal-datasets), and set mangal_dev_v2 as master
• Move generic API function (GET) to ZZZ.r to improve user experience

Hi,

I've just noticed a curious behavior of search_datasets(). When I tried to use purrr::map_ the code execution is way more slow, see below the result of this benchmark:

### Data

dataset_name <- c("fautin_1993", "ricciardi_2010", "edwards_1982")
 
#### First benchmark
tini <- Sys.time()

res_test_1 <- dataset_name %>%
  map_dfr(~search_datasets(search = x))
 
tfin <- Sys.time()

tfin - tini

Time difference of 3.554202 mins

### Second benchmark
tini <- Sys.time()

res_test_2 <- search_datasets(dataset_name[1])

Found 1 dataset(s) for query: fautin_1993

for(i in 2:length(dataset_name)){
  res_test_2 <- bind_rows(res_test_2, search_datasets(dataset_name[i]))
}

Found 1 dataset(s) for query: ricciardi_2010
Found 1 dataset(s) for query: edwards_1982

tfin <- Sys.time()

tfin - tini

Time difference of 2.091179 secs

I can't explain to myself this huge difference.

message() is called in many functions, I think it would make sense to add an argument verbose to suppress them if required.

Make sure that every entries are of the good type : use as.character or as.integer inside each POST_*() function.

Hi,

Maybe there were some recents change about db architecture, because list_datasets() is not working anymore.

When I tried to download a dataset (ie: kemp_1977) I got this:

> list_datasets("kemp_1977")

Found 1 dataset(s) for keywork: kemp_1977
Erreur : All columns in a tibble must be 1d or 2d objects:

• Column doi is NULL
• Column jstor is NULL
• Column pmid is NULL
• Column paper_url is NULL
  Call rlang::last_error() to see a backtrace

See the return of the db below.

Args

• doi
• query for full search

related to issue #27

If a function POST()failed, the length(httr::content) is greater than 1. But the printed error message for the user is print("attribute already in mangal") and it is not accurate.

Mangal db structure offers the opportunity to store traits taken on nodes (nodes table) and/or to store generic traits on taxa (taxonomy table). I plan to add a logical argument expand_trait (default: FALSE) to the function get_network_by_id() to add traits to mgNetwork (network$traits).

So far the current approach is to add data frame in column which I think is confusing for users.
This is the case for the networks column and the the additional appended when expand_node = TRUE. For the latter I propose something like :

if (expand_node) {
    tmp <- as.data.frame(get_singletons(endpoints()$node,
        interactions$node_from))
    names(tmp) <- paste0("node_from_", names(tmp))
    interactions <- cbind(tmp, interactions)
    #
    tmp <- as.data.frame(get_singletons(endpoints()$node,
        interactions$node_to))
    names(tmp) <- paste0("node_to_", names(tmp))
    interactions <- cbind(tmp, interactions)
 }

But there may be better option.

Hard to fix this one... don't know what it is.
@KevCaz , could you have a look if you have a moment.
Thanks !

> pkgdown::build_article("rmangal")
Reading 'vignettes/rmangal.Rmd'
Error in (function (classes, fdef, mtable)  : 
  unable to find an inherited method for function 'mapView' for signature '"tbl_df"'

Il faudrait faire un checkup pour être certain que les fonctions POST couvrent bien tout les types de messages d'erreur que l'API peut lancer

• POST_trait
• POST_line
• POST_table
• ...

Are raw and list useless in this context?

R> res4 <- search_datasets(query = "2011", output = "list")
Found 1 datasets
Error in datasets[i, "id"] : no 'dimnames' attribute for array
R> res4 <- search_datasets(query = "2011", output = "raw")
Found 1 datasets
Error in datasets[i, "id"] : no 'dimnames' attribute for array

Hi guys,

I am testing the new functions of rmangal. Right now I am trying to retrieve datasets whom I know the name. According to the documentation list_datasets() can accept arguments to pass at the get_gen() function.

So I want for exemple retrieve the dataset named kemp_1977, I try to do this:

list_datasets(query = paste0("name=", "kemp_1977"))

Or I try to do something like this :

list_datasets(query = list(name = "kemp_1997"))

Each time I got back the same error message

Error in get_gen(endpoints()$dataset, query = list(q = search), ...) :
argument formel "query" correspondant à plusieurs arguments fournis

Is it a bug or am I doing something wrong?

Test commented in #43

   Running the tests in ‘tests/testthat.R’ failed.
   Last 13 lines of output:
     Did you rename it, without setting st_geometry(obj) <- "newname"?
     1: search_networks(area) at testthat/test-search_networks.R:8
     2: sf::st_transform(polygon, crs = 4326)
     3: st_transform.sf(polygon, crs = 4326)
     4: st_transform(st_geometry(x), crs, ...)
     5: st_geometry(x)
     6: st_geometry.sf(x)
     7: stop("attr(obj, \"sf_column\") does not point to a geometry column.\nDid you rename it, without setting st_geometry(obj) <- \"newname\"?")
     
     ══ testthat results  ═══════════════════════════════════════════════════════════
     OK: 12 SKIPPED: 0 FAILED: 1
     1. Error: (unknown) (@test-search_networks.R#8) 
     
     Error: testthat unit tests failed
     Execution halted

Create https://mangal-wg.github.io/rmangal/ documentation website

reprex :

out  <- get_from_fkey_net(endpoints()$interaction, network_id = 19)

La fonction qui envoie la table ref n'a pas le bon URL. Aucune indication sur ce que serait le bon URL.

S'assurer que les fct GET utilisent aussi le tolower au bon endroit

https://github.com/mangal-wg/mangal-datasets/issues/4

Search network interactions based on type:

avail_type <- function() c(
  "competition",
  "amensalism",
  "neutralism",
  "commensalism",
  "mutualism",
  "parasitism",
  "predation",
  "herbivory",
  "symbiosis",
  "scavenger",
  "detritivore",
  "unspecified"
)

Attributes that are used by the api for the unique constrain should be in lower case

-> Add tolower(attribute) in injection fonctions

Args

• query: full search on node.original_name (provided by authors) and taxonomy.name (Taxonomy homogenized)
• bold: specific search on taxonomy.bold reliable to bold systems infra
• eol: specific search on taxonomy.eol reliable to EOL infra
• tsn: specific search on taxonomy.tsn reliable to ITIS infra
• eol: specific search on taxonomy.eol reliable to Encyclopedia of life infra

If I am correct, tibble is only used for its well-thought print method. A way of using it without depending on tibble is to set the class of data frames to c("tbl_df", "tbl", "data.frame") then if tibble is loaded the tibble's print method will be used! I got the hint form Noam Ross (ropensci/software-review#244).

• Travis badges, R dev status, appVeyor status
• Getting started (install procedure)
• Authentification mandatory for POST / Procedure to get ORCID token

Hi,

I have the same issue that #32 but with the search_networks() function.

e.g. the number of pages to be downloaded.

I had tested the mg_network() function before the pulled request #24 and it was working fine. Now I have updated my rmangal version, I got an error when I try to use this function :

mg_network(16) # id of kemp_1977 datatset's

API request failed: [404]
FALSE

It seems that the function is stuck in a infinite loop or something like this, because I need to abort the operation by hand. It is not a server issue beaucoup with my own function clementviolet/get_mangal/, I can access to this network and in Postman too.

List of comprehensive vignettes

1. How to list datasets?
2. How to retrieve networks from one and/or list of datasets?
3. How to retrieve all/specific networks?
4. Search over taxonomic name, retrieve networks which include this taxa
5. Search by location, used buffer, polygons to retrieve networks.
6. How to use the ids (bold, tsn, ncbi etc.) to get extra taxonomic informations using taxize?

Hi everyone!

I am writing my manuscript about my internship, and I would like to cite the rmangal package. I don't know how you want it to be cited, but I am making a proposal below.

@Manual{,
title = {rmangal},
author = {}, 
year = {2019},
note = {R package version 0.6.0},
url = {https://github.com/mangal-wg/rmangal}
}

I intentionally left blank the author field, because I don't know the order of the authors.

rbind.sf() is not exported by sf (you can access it with sf:::rbind.sf()) so :

purrr::reduce(purrr::map(responses,"body"), sf::rbind.sf)

(in api_utils.R) won't work:

R> sf::rbind.sf

Error: 'rbind.sf' is not an exported object from 'namespace:sf'

It looks like we can use the default rbind() (as explains here r-spatial/sf#92), worst case scenario we will have to copy paste the code of the function.

• Set CI/CD with travis
• Protect master branch
• Set/Init tests folder

nets <- search_networks("insect", output = "spatial")
names(nets)

We should rename the column as geom

1. If the geom.type is polygon, then only the first coordinates is returned.
   See get_network_by_id(1101)$network$geom.coordinates

2. If the id doesn't exist, the function return a 404 error which is right, but doesn't stop following API calls for nodes, edges, etc.

We need a function that will 'turn the page' when there is more than 100 entry for a GET request

Could be similiar to the V1 : https://github.com/mangal-wg/rmangal-v1/blob/master/R/pagerResources.R

Change the type of variable from FLOAT to STRING, so we can put qualitative environmental information as well.