epiverse-trace / sivirep Goto Github PK

Relevant function is here:

Can we avoid the eval(parse()) and use a design that would allow users to plug in their own data file of issues (excel or other)?

Comportamiento Esperado

Render PDF de la información de una enfermedad, año y departamento especificados.

Comportamiento Actual

Error

Pasos a Reproducir

secuencia de pasos que se requieren para poder reproducir el defecto. Si en la ejecución del escenario usted hizo uso de algún dato, debe detallar en los pasos, que dato uso y en donde lo uso.

1. Verificar que en el archivo PruebasPDF.Rmd (o cualquier nombre que se le haya dado al archivo) del template del reporte la línea 15 sea
   • \usepackage[spanish]{babel}
2. Crear el reporte al ejecutar:
   library(tinytex)
   rmarkdown::render("PruebasPDF/PruebasPDF.Rmd", params = list(
   nombre_evento = "Malaria",
   year = 2017,
   departmento = "Choco"
   ))
3. Esperar hasta el final de la ejecución, pues el error se obtiene al final.

Evidencias

! Package babel Error: Unknown option 'spanish'. Either you misspelled it
(babel) or the language definition file spanish.ldf
(babel) was not found.
(babel) There is a locale ini file for this language.
(babel) If it’s the main language, try adding `provide=*'
(babel) to the babel package options.

Error: LaTeX failed to compile PruebasPDF.tex. See https://yihui.org/tinytex/r/#debugging for debugging tips. See PruebasPDF.log for more info.
Además: There were 50 or more warnings (use warnings() to see the first 50)

Contexto (Ambiente de Ejecución)

En esta sección se debe detallar el contexto en el que fue ejecutado el escenario de prueba.

• Se probó en Windows 11 23H2
• R versión 4.3.2
• RStudio version 2023.12.1 Build 402

Información Adicional

Solución:
En la línea 15 del archivo PruebasPDF.Rmd donde inicialmente se tenía:

• \usepackage[spanish]{babel}
  reemplazar por:
• \usepackage[spanish, provide=*]{babel}

A pesar de no obtener error, se obtienen más de 50 warnings

Esto fue reportado por: @juanitaromerog

En la creación del mapa de malaria, el codigo de departamento no se ve reflejado en el mapa.

En este, aparece el número de orden en la que aparecen los departamento de la tabla de descripción.

Malaria and Mortalidad Materna

Warning | Unknown or uninitialised column: cod_eve

Most tidyverse functions use tidy evaluation, also called non-standard evaluation. This means that you don't use quotes to refer to something, even though this something is not an object in your R environment. For example:

# standard evaluation
library("ggplot2")

# non-standard evaluation. This should normally fail because we didn't define any object called "ggplot2". But it works due to some programming tricks inside the library() function
library(ggplot2)

In the case of dplyr, this takes the following form:

cars %>%
  # Note the absence of quotes around speed even though it's not a real object
  dplyr::filter(speed > 10)

The problem is that this will generate notes in R CMD check:

no visible binding for global variable ...

The proper solution, as documented in the vignette 'Programming with dplyr', and as documented by Maëlle Salmon in the RStudio community forum, is to use the .data prefix:

cars %>%
  dplyr::filter(.data$speed > 10)

dplyr/tidyverse and data.table has mostly the same overall set of features and it's generally not necessary to depend on both.

To improve code maintainability and sustainability by avoiding unnecessary dependencies, it would be good to decide about using just one of them.

I suspect the dependency on data.table (only used twice: https://github.com/search?q=repo%3Aepiverse-trace%2Fsivirep+data.table%3A%3A&type=code) is currently weaker than the dependency on dplyr so it's probably easier to remove it.

The function agrupar_mpio generates an error when words with special characters are sent in the dpto parameter:

agrupar_mpio( data_event, dpto = "Choco", nomb_col = "cod_mun_o", porcentaje = FALSE )

Reported by: @juanitaromerog

Geral. Podrías por favor agregar una columna de cod_eve a la funcion lista_eventos?
No es urgente, pero es bastante útil tenerla allí
Gracias!

As part of #149 CRAN Policies indicate that:

Packages which use Internet resources should fail gracefully with an informative message if the resource is not available or has changed (and not give a check warning nor error).

All connections to SIVIGILA servers should address the above

Si es posible incluir una función en donde se puedan obtener tiempos entre las fechas, por ejemplo, tiempo promedio entre fecha de inicio de síntomas y fecha de notificación, fecha de inicio de síntomas y fecha de defunción y fecha de inicio de síntomas y fecha de consulta.

Aporte Grupo 4 Sivirep.

The function geo_filtro generates an error when words with special characters are sent in the dpto and mpio parameters:

geo_filtro(data_event = data_limpia, dpto = "PASTO", mpio = "NARIÑO")

Reported by: @juanitaromerog

sivirep currently provides a very nice looking report and it would be good to already display it to potential users when they discover the package.

A good way to do this could be to add a copy of the pdf on the pkgdown website or/and a screenshot in the README.

I can try to submit a pull request if you agree with this idea but you're not sure how to implement it.

Not 100% sure but it looks like you have hardcoded a path, which means that malaria data will be fetched in import_data_endemic_channel(), no matter what the user passes as disease_name:

spatial_dept_dist <- group_dept(disease_data = clean_disease_data)
is giving as output just id and casos

spatial_dept_dist
# A tibble: 16 × 2
   id    casos
   <chr> <int>
 1 1         1
 2 5         5
 3 8         4
 4 15       19

can you please add name of department, to the output?
With this feature, the output becomes helpful for making tables. A similar feature for municipalities is needed. Thanks!

The examples in cleaning_data.R all fail because the data you are cleaning has not been defined beforehand. For example, in:

you need to define deptos_data.

Hola Geral

• [x ] I have the most recent version of sivirep and R
• [x ] I have found a bug

Creo que he encontrado un bug pero no estoy segura cómo funciona porque el problema sólo ocurre con Cali en esta serie de ejemplos. Cuando uso el geofiltro para municipio Cali, me sale un total de 17 casos cuando deberían ser 13376. Pero lo probé para otros municipios del Valle y de Amazonas y no sale ese problema. No sé si sea algo con el nombre CALI. y tal vez pueda ocurrir con otros.

Dejo el código reproducible

library(sivirep)
dat <- import_data_event(year = 2020, nombre_event = "Dengue")
dat_clean <- limpiar_data_sivigila(data_event= dat)
dat_valle <- geo_filtro(data_event = dat_clean, dpto = "VALLE")
dat_cali  <- geo_filtro(data_event= dat_clean, dpto = "VALLE", mpio = "CALI")
dist_esp_mun <- agrupar_mpio(data_event = dat_valle, dpto = "VALLE")

NROW(dat_cali) # 17 casos 
dist_esp_mun %>% filter(municipio_ocurrencia == "CALI") # 13376 


dat_buga  <- geo_filtro(data_event= dat_clean, dpto = "VALLE", mpio = "BUGA")
NROW(dat_buga) # 1312 casos
dist_esp_mun %>% filter(municipio_ocurrencia == "BUGA") # 1312 casos

dat_amazonas <- geo_filtro(data_event = dat_clean, dpto = "AMAZONAS")
dat_leticia  <- geo_filtro(data_event= dat_clean, dpto = "AMAZONAS", mpio = "LETICIA")
dist_esp_amazonas_mun <- agrupar_mpio(data_event = dat_amazonas, dpto = "AMAZONAS")
NROW(dat_leticia) # 176 casos
dist_esp_amazonas_mun %>% filter(municipio_ocurrencia == "LETICIA") # 176 casos

¡Mil gracias!

The example dataset dengue2020 is currently too large. It's essential to reduce its size for both performance optimization and to meet CRAN's submission requirements.

sivirep currently provides many utility function to manipulate data provided by the SIVIGILA source.

As a result, the pkgdown reference index can be a bit overwhelming. It would be good to group the functions into logical sections, as ggplot2 for example is doing.

Your good design work makes it particularly easy here since you have use consistent function prefixes and names moulds throughout the package so you can just use selectors such as starts_with().

Related pkgdown documentation: https://pkgdown.r-lib.org/reference/build_reference.html#reference-index

I have not reviewed the package code or design but upon a cursory look, here are the points that almost surely will cause issues with CRAN:

• the package is too large (>= 5MB):
  • some data is duplicated across data and inst/extdata, can this be fixed?
  • can a subset of the dengue2020 dataset be used in examples? Examples are just here to demonstrate the function usage, it's not so important if the result is not necessarily meaningful from a scientific point of view. Vignettes are here to fill this gap.
  • the getting_started.Rmd and sivirep.Rmd vignettes seem mostly duplicated (both leading to an almost 50MB HTML file)
  • I would convert the gif to webm, which is widely supported by all web browsers now (source): I achieved a 20x file size reduction doing this (with ffmpeg). Please let me know if you'd like me to submit a PR with this change
• examples take too long
  • this should be improved automatically if you follow the smaller example dataset recommendation above
  • longer examples should be wrapped in \donttest{}
• the package is in Spanish. This is not a problem in itself but it's uncommon enough that CRAN is likely to raise an eyebrow over this. I would pre-empt any comments by explaining in cran-comments.md to scope of the package, why it has to be in Spanish, and how you have tried to follow recommendation from the official 'Writing R Extensions' guide: https://cran.r-project.org/doc/manuals/R-exts.html#The-DESCRIPTION-file-1

We want to give the user the option of active or deactive the different sections of the "Reporte Basico" template

It is recommended to organize municipalities by number of cases to improve understanding and decision making.

Respecto a la generación grafica, cuando se seleccione solo un municipio se podria visualizar todo el departamento y sombrear especificamente el municipio, para no ser redundante en graficas de este tipo:

Por otra parte seria interesante ver análisis comparativos entre departamentos al permitir seleccionar mas de una opción en esta variable.

Isuue relate #146

The order of variables in plotting functions affect the visualizations. These are some examples:

plot_edad_sex(data_agrupada, nomb_cols = c("edad", "sexo"), fuente_data = NULL)
plot_sex_semanaepi(data_agrupada,nomb_cols = c("sexo", "semana"),fuente_data = "SIVIREP")

Reported by: @juanitaromerog

The function agrupar_edad_sex doesn't display anything when the parameter porcentaje = TRUE is passed.

agrupar_edad_sex( data_event_limp, nomb_cols = c("edad", "sexo"), porcentaje = TRUE, interval_edad = 15)

Reported by: @juanitaromerog

When generating a custom report from "Reporte básico" template. When exporting to html, the document does not show any numbering on tables nor figures

Hi Gerald! Just before I forget, this is the typical visualisation for the endemic channel from Sivigila. Perhaps we need to work on this visualisation as a new feature for multi-year reports for both national and subnational. For the national, it would be ideal to have the plot for 32 departments at once.

Dengue data from 2016 have a different format than other years

Es una buena aproximación al análisis exploratorio de las bases publicas de Sivigila, algunas mejoras de "forma" incluyen:

Los ejes de edad son "grupos de edad" no edades simples.
Nombrar las gráficas en la parte superior.
La gráfica de fecha de notificación esta repetida con la gráfica de fecha por inicio de síntomas.

Un cambio importante es que Dengue y Malaria se debe calcular sobre la población a riesgo que provee Ministerio de Salud y no sobre población total de las proyecciones DANE.

Muchas gracias por el trabajo porque permite automatizar el proceso de análisis descriptivo de históricos.