updating readme [skip ci]

Dockerfile

@@ -1,8 +1,8 @@
 FROM bentoml/model-server:0.11.0-py37
 MAINTAINER ersilia
-
-RUN conda install -c conda-forge rdkit=2019.03.01
-RUN pip install signaturizer==1.1.10
+ 
+RUN pip install rdkit-pypi
+RUN pip install signaturizer==1.1.10 
 
 WORKDIR /repo
-COPY . /repo
+COPY . /repo

README.md

@@ -25,6 +25,8 @@ A set of 25 Chemical Checker bioactivity signatures (including 2D & 3D fingerpri
 
 ## Ersilia model URLs
 * [GitHub](https://github.com/ersilia-os/eos4u6p)
+* [AWS S3](https://ersilia-models-zipped.s3.eu-central-1.amazonaws.com/eos4u6p.zip)
+* [DockerHub](https://hub.docker.com/r/ersiliaos/eos4u6p) (AMD64, ARM64)
 
 ## Citation
 

metadata.json

@@ -28,5 +28,11 @@
     "Publication": "https://www.nature.com/articles/s41467-021-24150-4",
     "Source Code": "http://gitlabsbnb.irbbarcelona.org/packages/signaturizer",
     "License": "MIT",
-    "Contributor": "miquelduranfrigola"
+    "Contributor": "miquelduranfrigola",
+    "S3": "https://ersilia-models-zipped.s3.eu-central-1.amazonaws.com/eos4u6p.zip",
+    "DockerHub": "https://hub.docker.com/r/ersiliaos/eos4u6p",
+    "Docker Architecture": [
+        "AMD64",
+        "ARM64"
+    ]
 }

model/framework/run.sh

@@ -0,0 +1 @@
+python $1/predict.py $2 $3

src/service.py

@@ -25,9 +25,9 @@ def load_model(framework_dir, checkpoints_dir):
 
 class Model(object):
     def __init__(self):
-        self.DATA_FILE = "data.csv"
-        self.PRED_FILE = "pred.json"
-        self.RUN_FILE = "run.sh"
+        self.DATA_FILE = "_data.csv"
+        self.OUTPUT_FILE = "_output.csv"
+        self.RUN_FILE = "_run.sh"
         self.LOG_FILE = "run.log"
 
     def load(self, framework_dir, checkpoints_dir):
@@ -40,22 +40,22 @@ def set_checkpoints_dir(self, dest):
     def set_framework_dir(self, dest):
         self.framework_dir = os.path.abspath(dest)
 
-    def predict(self, smiles_list):
-        tmp_folder = tempfile.mkdtemp()
+    def run(self, input_list):
+        tmp_folder = tempfile.mkdtemp(prefix="eos-")
         data_file = os.path.join(tmp_folder, self.DATA_FILE)
-        pred_file = os.path.join(tmp_folder, self.PRED_FILE)
+        output_file = os.path.join(tmp_folder, self.OUTPUT_FILE)
         log_file = os.path.join(tmp_folder, self.LOG_FILE)
         with open(data_file, "w") as f:
-            f.write("smiles"+os.linesep)
-            for smiles in smiles_list:
-                f.write(smiles + os.linesep)
+            f.write("input" + os.linesep)
+            for inp in input_list:
+                f.write(inp + os.linesep)
         run_file = os.path.join(tmp_folder, self.RUN_FILE)
         with open(run_file, "w") as f:
             lines = [
-                "python {0}/predict.py {1} {2}".format(
+                "bash {0}/run.sh {0} {1} {2}".format(
                     self.framework_dir,
                     data_file,
-                    pred_file
+                    output_file
                 )
             ]
             f.write(os.linesep.join(lines))
@@ -64,8 +64,9 @@ def predict(self, smiles_list):
             subprocess.Popen(
                 cmd, stdout=fp, stderr=fp, shell=True, env=os.environ
             ).wait()
-        with open(pred_file, "r") as f:
+        with open(output_file, "r") as f:
             result = json.load(f)
+        shutil.rmtree(tmp_folder)    
         return result
 
 
@@ -119,8 +120,8 @@ def save(self, dst):
 @artifacts([Artifact("model")])
 class Service(BentoService):
     @api(input=JsonInput(), batch=True)
-    def predict(self, input: List[JsonSerializable]):
+    def run(self, input: List[JsonSerializable]):
         input = input[0]
-        smiles_list = [inp["input"] for inp in input]
-        output = self.artifacts.model.predict(smiles_list)
+        input_list = [inp["input"] for inp in input]
+        output = self.artifacts.model.run(input_list)
         return [output]