Use reticulate instead of PythonInR, alanwood has moved

Because the pesticide compendium at alanwood.net has moved to bcpc.org,
we need to depend on a development branch of webchem at the moment.

7 files changed, 118 insertions, 84 deletions

diff --git a/.travis.yml b/.travis.yml
index 712829e..841fe6d 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -5,5 +5,6 @@ r:
   - devel
 github_packages:
   - r-lib/covr
+  - jranke/webchem@bcpc
 after_success:
   - Rscript -e 'covr::codecov()'
diff --git a/ChangeLog b/ChangeLog
index 011b28d..69b201d 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,12 @@
+commit 898d598bcbb022ac767e0105e868798130f2c74c
+Author: Johannes Ranke <jranke@uni-bremen.de>
+Date:   2020-10-13 15:10:03 +0200
+
+    Adapt to changes in webchem
+    
+    Unfortunately, tests only pass sometimes, it seems PubChem is not always
+    responding
+
 commit 6937e41d317b953d8246203814500166dbe89470
 Author: Johannes Ranke <jranke@uni-bremen.de>
 Date:   2020-10-13 11:26:03 +0200
diff --git a/DESCRIPTION b/DESCRIPTION
index a00e9a4..041999c 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -1,13 +1,13 @@
 Package: chents
 Type: Package
 Title: Chemical Entities as R Objects
-Version: 0.2-9
-Date: 2020-10-13
+Version: 0.3.1
+Date: 2021-11-18
 Authors@R: c(person("Johannes", "Ranke", role = c("aut", "cre", "cph"),
                     email = "jranke@uni-bremen.de"))
 Description: Utilities for dealing with chemical entities and associated
     data as R objects. If Python and RDKit (> 2015.03) are installed and
-    configured for use with 'PythonInR', some basic chemoinformatics functions
+    configured for use with 'reticulate', some basic chemoinformatics functions
     like the calculation of molecular weight and plotting of chemical
     structures in R graphics are available. For plotting, you need
     grConvert (https://sjp.co.nz/projects/grconvert) which is not on CRAN.
@@ -20,7 +20,7 @@ Imports:
 Suggests:
     knitr,
     testthat,
-    PythonInR,
+    reticulate,
     covr,
     devEMF
 License: GPL
@@ -28,4 +28,4 @@ LazyLoad: yes
 LazyData: yes
 Encoding: UTF-8
 URL: https://github.com/jranke/chents
-RoxygenNote: 7.1.1.9000
+RoxygenNote: 7.1.2
diff --git a/R/chent.R b/R/chent.R
index b2c2e4a..d988e97 100644
--- a/R/chent.R
+++ b/R/chent.R
@@ -1,4 +1,4 @@
-# Copyright (C) 2016-2019  Johannes Ranke
+# Copyright (C) 2016-2021  Johannes Ranke
 # Contact: jranke@uni-bremen.de
 # This file is part of the R package chents
 
@@ -19,7 +19,7 @@
 #'
 #' The class is initialised with an identifier. Chemical information is retrieved from
 #' the internet. Additionally, it can be generated using RDKit if RDKit and its
-#' python bindings are installed and configured for use with PythonInR.
+#' python bindings are installed.
 #'
 #' @export
 #' @format An \code{\link{R6Class}} generator object
@@ -33,8 +33,8 @@
 #' @field smiles SMILES code, with attribute 'source'
 #' @field mw Molecular weight, with attribute 'source'
 #' @field pubchem List of information retreived from PubChem
-#' @field rdkit List of information obtained with RDKit, if installed and
-#'   configured for use with PythonInR
+#' @field rdkit List of information obtained with RDKit
+#' @field mol <rdkit.Chem.rdchem.Mol> object
 #' @field svg SVG code
 #' @field Picture Graph as a \code{\link{picture}} object obtained using grImport
 #' @field Pict_font_size Font size as extracted from the intermediate PostScript file
@@ -47,6 +47,18 @@
 #' @field soil_sorption Dataframe of soil sorption data
 #' @field PUF Plant uptake factor
 #' @keywords data
+#' @examples
+#' oct <- chent$new("1-octanol", smiles = "CCCCCCCCO")
+#' print(oct)
+#' if (!is.null(oct$Picture)) {
+#'   plot(oct)
+#' }
+#'
+#' caffeine <- chent$new("caffeine")
+#' print(caffeine)
+#' if (!is.null(caffeine$Picture)) {
+#'   plot(caffeine)
+#' }
 
 chent <- R6Class("chent",
   public <- list(
@@ -56,6 +68,7 @@ chent <- R6Class("chent",
     mw = NULL,
     pubchem = NULL,
     rdkit = NULL,
+    mol = NULL,
     svg = NULL,
     Picture = NULL,
     Pict_font_size = NULL,
@@ -86,7 +99,7 @@ chent <- R6Class("chent",
       }
 
       if (rdkit) {
-        if(rdkit_available()) {
+        if(rdkit_available) {
           if (is.null(self$smiles)) {
             message("RDKit would need a SMILES code")
           } else {
@@ -98,7 +111,7 @@ chent <- R6Class("chent",
             attr(self$mw, "source") <- "rdkit"
           }
         } else {
-          message("RDKit is not available via PythonInR")
+          message("RDKit is not available")
         }
       }
 
@@ -158,13 +171,12 @@ chent <- R6Class("chent",
       }
     },
     get_rdkit = function(template = NULL) {
-      if(!rdkit_available()) {
-        stop("RDKit is not available via PythonInR")
+      if(!rdkit_available) {
+        stop("RDKit is not available")
       }
       self$rdkit <- list()
-      PythonInR::pyImport("Descriptors", from = "rdkit.Chem")
-      PythonInR::pyExec(paste0("mol = Chem.MolFromSmiles('", self$smiles[1], "')"))
-      self$rdkit$mw <- PythonInR::pyExecg("mw = Descriptors.MolWt(mol)", "mw")
+      self$mol <- rdkit_module$Chem$MolFromSmiles(self$smiles[1])
+      self$rdkit$mw <- rdkit_module$Chem$Descriptors$MolWt(self$mol)
       if (!is.null(self$mw)) {
         if (round(self$rdkit$mw, 1) != round(self$mw, 1)) {
           message("RDKit mw is ", self$rdkit$mw)
@@ -173,20 +185,16 @@ chent <- R6Class("chent",
       }
 
       # Create an SVG representation
-      PythonInR::pyImport("Draw", from = "rdkit.Chem")
-      PythonInR::pyImport("rdMolDraw2D", from = "rdkit.Chem.Draw")
-      PythonInR::pyImport("rdDepictor", from = "rdkit.Chem")
-      PythonInR::pyExec("rdDepictor.Compute2DCoords(mol)")
+      rdkit_module$Chem$rdDepictor$Compute2DCoords(self$mol)
       if (!is.null(template)) {
-        PythonInR::pyImport("AllChem", from = "rdkit.Chem")
-        PythonInR::pyExec(paste0("template = Chem.MolFromSmiles('", template, "')"))
-        PythonInR::pyExec("AllChem.Compute2DCoords(template)")
-        PythonInR::pyExec("AllChem.GenerateDepictionMatching2DStructure(mol, template)")
+        rdkit_template <- rdkit_module$Chem$MolFromSmiles(template)
+        rdkit_module$Chem$rdDepictor$Compute2DCoords(template)
+        rdkit$Chem$AllChem$GenerateDepictionMatching2DStructure(self$mol, template)
       }
-      PythonInR::pyExec("d2d = rdMolDraw2D.MolDraw2DSVG(400,500)")
-      PythonInR::pyExec("d2d.DrawMolecule(mol)")
-      PythonInR::pyExec("d2d.FinishDrawing()")
-      self$svg <- PythonInR::pyGet("d2d.GetDrawingText()")
+      d2d <- rdkit_module$Chem$Draw$rdMolDraw2D$MolDraw2DSVG(400L, 400L)
+      d2d$DrawMolecule(self$mol)
+      d2d$FinishDrawing()
+      self$svg <- d2d$GetDrawingText()
       svgfile <- tempfile(fileext = ".svg")
       psfile <- tempfile(fileext = ".ps")
       writeLines(self$svg, svgfile)
@@ -446,19 +454,14 @@ print.chent = function(x, ...) {
 draw_svg.chent = function(x, width = 300, height = 150,
                           filename = paste0(names(x$identifier), ".svg"),
                           subdir = "svg") {
-  if (!PythonInR::pyIsConnected()) {
-    PythonInR::pyConnect()
-  }
-  try_rdkit <- try(PythonInR::pyImport("Chem", from = "rdkit"))
-  if (inherits(try_rdkit, "try-error")) {
-    message("Could not import RDKit in Python session")
+  if (!rdkit_available) {
+    stop("RDkit is not available via reticulate")
   } else {
     if (!dir.exists(subdir)) dir.create(subdir)
-    PythonInR::pyExec(paste0("mol = Chem.MolFromSmiles('", x$smiles, "')"))
-    PythonInR::pyImport("Draw", from = "rdkit.Chem")
-    cmd <- paste0("Draw.MolToFile(mol, '", file.path(subdir, filename),
-                  "', size = (", width, ", ", height, "))")
-    PythonInR::pyExec(cmd)
+    mol <- rdkit_module$Chem$MolFromSmiles(x$smiles)
+
+    rdkit_module$Chem$Draw$MolToFile(mol, file.path(subdir, filename),
+      size = c(as.integer(width), as.integer(height)))
   }
 }
 
@@ -482,19 +485,25 @@ plot.chent = function(x, ...) {
 #' @importFrom R6 R6Class
 #' @export
 #' @format An \code{\link{R6Class}} generator object
-#' @field iso ISO common name according to ISO 1750 as retreived from www.alanwood.net/pesticides
-#' @field alanwood List of information retreived from www.alanwood.net/pesticides
+#' @field iso ISO common name according to ISO 1750 as retreived from pesticidecompendium.bcpc.org
+#' @field bcpc List of information retrieved from pesticidecompendium.bcpc.org
 #' @keywords data
+#' @examples
+#' atr <- pai$new("atrazine")
+#' print(atr)
+#' if (!is.null(atr$Picture)) {
+#'   plot(atr)
+#' }
 
 pai <- R6Class("pai",
   inherit = chent,
   public <- list(
     iso = NULL,
-    alanwood = NULL,
+    bcpc = NULL,
     initialize = function(iso, identifier = iso,
                           smiles = NULL, smiles_source = 'user',
                           inchikey = NULL, inchikey_source = 'user',
-                          alanwood = TRUE,
+                          bcpc = TRUE,
                           pubchem = TRUE, pubchem_from = 'auto',
                           rdkit = TRUE, template = NULL,
                           chyaml = TRUE)
@@ -505,28 +514,28 @@ pai <- R6Class("pai",
         attr(self$inchikey, "source") <- "user"
       }
 
-      if (!missing(iso) & alanwood) {
-        message("alanwood.net:")
-        aw_result = webchem::aw_query(identifier, from = "name")
+      if (!missing(iso) & bcpc) {
+        message("BCPC:")
+        bcpc_result = webchem::bcpc_query(identifier, from = "name")
 
         # Use first element of list, as we passed a query of length one
-        if (is.na(aw_result[[1]][1])) {
-          message("Common name ", identifier, " is not known at www.alanwood.net, trying PubChem")
+        if (is.na(bcpc_result[[1]][1])) {
+          message("Common name ", identifier, " is not known at the BCPC compendium, trying PubChem")
         } else {
-          self$alanwood = aw_result[[1]]
-          self$iso = self$alanwood$cname
-          attr(self$iso, "source") <- "alanwood"
-          attr(self$iso, "status") <- self$alanwood$status
-          aw_ik = self$alanwood$inchikey
-          if (length(aw_ik) == 1 && nchar(aw_ik) == 27 && !is.na(aw_ik)) {
+          self$bcpc = bcpc_result[[1]]
+          self$iso = self$bcpc$cname
+          attr(self$iso, "source") <- "bcpc"
+          attr(self$iso, "status") <- self$bcpc$status
+          bcpc_ik = self$bcpc$inchikey
+          if (length(bcpc_ik) == 1 && nchar(bcpc_ik) == 27 && !is.na(bcpc_ik)) {
             if (is.null(self$inchikey)) {
-              self$inchikey = self$alanwood$inchikey
-              attr(self$inchikey, "source") <- "alanwood"
+              self$inchikey = self$bcpc$inchikey
+              attr(self$inchikey, "source") <- "bcpc"
             } else {
-              if (aw_ik == self$inchikey) {
-                attr(self$inchikey, "source") = c(attr(self$inchikey, "source"), "alanwood")
+              if (bcpc_ik == self$inchikey) {
+                attr(self$inchikey, "source") = c(attr(self$inchikey, "source"), "bcpc")
               } else {
-                warning("InChIKey ", self$inchikey, " differs from ", aw_ik, " obtained from alanwood.net")
+                warning("InChIKey ", self$inchikey, " differs from ", bcpc_ik, " obtained from bcpc.org")
               }
             }
           }
@@ -608,23 +617,4 @@ pp <- R6Class("pp",
   )
 )
 
-rdkit_available <- function()
-{
-  if(requireNamespace("PythonInR", quietly = TRUE)) {
-    if (!PythonInR::pyIsConnected()) {
-      PythonInR::pyConnect()
-    }
-    sink(tempfile())
-    try_rdkit <- try(PythonInR::pyImport("Chem", from = "rdkit"),
-                     silent = TRUE)
-    sink()
-    if (inherits(try_rdkit, "try-error")) {
-      return(FALSE)
-    } else {
-      return(TRUE)
-    }
-  } else {
-    return(FALSE)
-  }
-}
 # vim: set ts=2 sw=2 expandtab:
diff --git a/R/zzz.R b/R/zzz.R
new file mode 100644
index 0000000..19ca10d
--- /dev/null
+++ b/R/zzz.R
@@ -0,0 +1,13 @@
+.onLoad = function(libname, pkgname) {
+  rdkit_available <- FALSE
+  if(requireNamespace("reticulate", quietly = TRUE)) {
+    rdkit_module <- try(reticulate::import("rdkit"))
+    if (!inherits(rdkit_module, "try-error")) {
+      rdkit_available <- TRUE
+    }
+  }
+  assign('rdkit_available', rdkit_available, envir = topenv())
+  if (rdkit_available) {
+    assign('rdkit_module', rdkit_module, envir = topenv())
+  }
+}
diff --git a/man/chent.Rd b/man/chent.Rd
index a1a5fe8..d94452e 100644
--- a/man/chent.Rd
+++ b/man/chent.Rd
@@ -9,7 +9,20 @@ An \code{\link{R6Class}} generator object
 \description{
 The class is initialised with an identifier. Chemical information is retrieved from
 the internet. Additionally, it can be generated using RDKit if RDKit and its
-python bindings are installed and configured for use with PythonInR.
+python bindings are installed.
+}
+\examples{
+oct <- chent$new("1-octanol", smiles = "CCCCCCCCO")
+print(oct)
+if (!is.null(oct$Picture)) {
+  plot(oct)
+}
+
+caffeine <- chent$new("caffeine")
+print(caffeine)
+if (!is.null(caffeine$Picture)) {
+  plot(caffeine)
+}
 }
 \keyword{data}
 \section{Public fields}{
@@ -25,8 +38,9 @@ python bindings are installed and configured for use with PythonInR.
 
 \item{\code{pubchem}}{List of information retreived from PubChem}
 
-\item{\code{rdkit}}{List of information obtained with RDKit, if installed and
-configured for use with PythonInR}
+\item{\code{rdkit}}{List of information obtained with RDKit}
+
+\item{\code{mol}}{<rdkit.Chem.rdchem.Mol> object}
 
 \item{\code{svg}}{SVG code}
 
diff --git a/man/pai.Rd b/man/pai.Rd
index 6a243ab..c27a8f5 100644
--- a/man/pai.Rd
+++ b/man/pai.Rd
@@ -11,6 +11,13 @@ An \code{\link{R6Class}} generator object
 The class is initialised with an identifier which is generally an ISO common name.
 Additional chemical information is retrieved from the internet if available.
 }
+\examples{
+atr <- pai$new("atrazine")
+print(atr)
+if (!is.null(atr$Picture)) {
+  plot(atr)
+}
+}
 \keyword{data}
 \section{Super class}{
 \code{\link[chents:chent]{chents::chent}} -> \code{pai}
@@ -18,9 +25,9 @@ Additional chemical information is retrieved from the internet if available.
 \section{Public fields}{
 \if{html}{\out{<div class="r6-fields">}}
 \describe{
-\item{\code{iso}}{ISO common name according to ISO 1750 as retreived from www.alanwood.net/pesticides}
+\item{\code{iso}}{ISO common name according to ISO 1750 as retreived from pesticidecompendium.bcpc.org}
 
-\item{\code{alanwood}}{List of information retreived from www.alanwood.net/pesticides}
+\item{\code{bcpc}}{List of information retrieved from pesticidecompendium.bcpc.org}
 }
 \if{html}{\out{</div>}}
 }
@@ -64,7 +71,7 @@ Additional chemical information is retrieved from the internet if available.
   smiles_source = "user",
   inchikey = NULL,
   inchikey_source = "user",
-  alanwood = TRUE,
+  bcpc = TRUE,
   pubchem = TRUE,
   pubchem_from = "auto",
   rdkit = TRUE,