Merge pull request #19 from thobalose/dev

Added phenotype and antibiotic options, and updated CLI setup.

.travis.yml

@@ -26,7 +26,7 @@ install:
 
 before_script:
   - sudo apt-get update
-  - wget https://repo.continuum.io/miniconda/Miniconda3-4.3.31-Linux-x86_64.sh -O miniconda.sh
+  - wget https://repo.continuum.io/miniconda/Miniconda3-4.6.14-Linux-x86_64.sh -O miniconda.sh
   - bash miniconda.sh -b -p $HOME/miniconda
   - export PATH="$HOME/miniconda/bin:$PATH"
   - hash -r
@@ -36,6 +36,7 @@ before_script:
   - conda config --add channels conda-forge
   - conda config --add channels bioconda
   - conda config --add channels https://conda.anaconda.org/thoba
+  - conda update --all
   - conda info -a
   - conda install conda-build anaconda-client
 

Pipfile

@@ -12,6 +12,9 @@ verify_ssl = true
 vcf2neo = {editable = true,path = "."}
 autopep8 = "*"
 pylint = "*"
+pydocstyle = "*"
+pep8 = "*"
+flake8 = "*"
 
 [packages]
 beautifulsoup4 = "==4.5.3"
@@ -38,7 +41,7 @@ tqdm = "==4.10.0"
 wrapt = "==1.10.8"
 xmltodict = "==0.10.2"
 coverage = "*"
-combattbmodel = "*"
+combattbmodel = "==0.0.8"
 
 [requires]
 python_version = "3.7"

README.md

@@ -52,10 +52,20 @@ You change the default database location (`localhost`) by setting the
 `DATABASE_URL` environment variable to `remote`.
 
 ```sh
-$ vcf2neo --help
-Usage: vcf2neo [OPTIONS] COMMAND [ARGS]...
+$ vcf2neo load_vcf --help
+Usage: vcf2neo load_vcf [OPTIONS] VCF_DIR [OWNER] [HISTORY_ID] [OUTPUT_DIR]
+
+  Load SnpEff annotated VCF files to genes and drugs in NeoDb.
+
+Options:
+  -p, --phenotype [XDR|MDR|SUSCEPTIBLE|UNKNOWN]
+                                  Specify phenotype.  [required]
+  -a, --antibiotic TEXT           Specify antibiotic. E.g. Rifampicin
+                                  [required]
+  --help                          Show this message and exit.
+
 ...
-$ vcf2neo load_vcf PATH/TO/VCF_DIR
+$ vcf2neo load_vcf -p UNKNOWN -a UNKNOWN PATH/TO/VCF_DIR
 ```
 
 **Exploring variant data**:

dc/Dockerfile

@@ -3,18 +3,17 @@ LABEL Maintainer="thoba@sanbi.ac.za"
 
 RUN apk update \
     && apk upgrade \
-    && mkdir /data \
-    && apk add --no-cache ca-certificates openssl curl \
-    && update-ca-certificates
+    && mkdir -p /data/guides \
+    && apk add --no-cache openssl wget
 
-ENV DB_VERSION="3057235" \
-    # ZENODO_URL="https://zenodo.org/record/${DB_VERSION}/files/neodb-db-data.tar.bz2" \
-    FIGSHARE_URL="https://ndownloader.figshare.com/files/15377531"
+ENV DB_VERSION="3258008"
 
-RUN wget "${FIGSHARE_URL}" \
+RUN wget "https://zenodo.org/record/${DB_VERSION}/files/neodb-db-data.tar.bz2" \
     -O neodb_db_data.tar.bz2 \
     && tar xvfj neodb_db_data.tar.bz2 \
     && rm -rf neodb_db_data.tar.bz2 \
-    && cp -r databases /data/
+    && cp -r databases /data/ \
+    && wget https://raw.githubusercontent.com/COMBAT-TB/combat-tb-neodb/master/dc/guides/combattb_neodb.html \
+    -P /data/guides/
 
 VOLUME /data

docker-compose.yml

@@ -15,6 +15,7 @@ services:
       NEO4J_ACCEPT_LICENSE_AGREEMENT: 'yes'
       NEO4J_dbms_allow__format__migration: 'true'
       NEO4J_dbms_allow__upgrade: 'true'
+      NEO4J_browser_post__connect__cmd: "config; play http://localhost:7474/guides/combattb_neodb.html"
     ports:
       - "7687:7687"
       - "7474:7474"

setup.py

@@ -4,7 +4,7 @@
 
 setup(
     name='vcf2neo',
-    version='0.0.8',
+    version='0.0.9',
     url='https://github.com/COMBAT-TB/vcf2neo',
     bugtrack_url='https://github.com/COMBAT-TB/vcf2neo/issues',
     description='Parses SnpEff annotated VCF files and builds a graph '

test/test_cli.py

@@ -20,5 +20,8 @@ def cli_runner():
 
 
 def test_load_vcf(cli_runner):
-    result = cli_runner.invoke(load_vcf, [TEST_DATA_DIR])
+    result = cli_runner.invoke(
+        load_vcf,
+        ['-p', 'MDR', '-a', 'Rifampicin', '-a', 'Isoniazid', TEST_DATA_DIR]
+    )
     assert result.exit_code == 0

vcf2neo/cli.py

@@ -10,6 +10,8 @@
 from vcf2neo.docker import Docker
 from vcf2neo.vcfproc import process_vcf_files
 
+HOST = os.environ.get("DATABASE_URL", "localhost")
+
 
 @click.group()
 def cli():
@@ -19,45 +21,49 @@ def cli():
     pass
 
 
-try:
-    # Python 2
-    u_str = unicode
-except NameError:
-    # Python 3
-    u_str = str
-
-
 @cli.command()
 @click.argument('vcf_dir', type=click.Path(exists=True, dir_okay=True),
                 required=True)
-@click.argument('owner', type=u_str, required=False)
-@click.argument('history_id', type=u_str, required=False)
-@click.argument('output_dir', type=click.Path(exists=True, dir_okay=True),
-                required=False)
-@click.option('-d/-D', default=False, help='Run Neo4j docker container.')
-def load_vcf(vcf_dir, owner, history_id, d, output_dir=None):
+@click.option('--owner', required=True, default=os.environ.get('USER', ''),
+              show_default='Current $USER',  help='Specify owner.')
+# @click.argument('history_id', type=u_str, required=False)
+# @click.argument('output_dir', type=click.Path(exists=True, dir_okay=True),
+#                 required=False)
+# @click.option('--docker/--no-docker', default=False,
+#               help='Run Combat-TB-NeoDB container.')
+@click.option('--phenotype', '-p', required=True,
+              type=click.Choice(['XDR', 'MDR', 'SUSCEPTIBLE', 'UNKNOWN']),
+              help='Specify phenotype.')
+@click.option('--antibiotic', '-a', multiple=True, required=True,
+              help='Specify antibiotic. E.g. Rifampicin')
+def load_vcf(vcf_dir, owner, phenotype=None, antibiotic=None):
     """
+
     Load SnpEff annotated VCF files to genes and drugs in NeoDb.
     """
-    docker = None
-    if d:
+
+    # TODO: Look into docker implemantation
+    container, docker, output_dir, history_id = None, None, None, None
+    if docker:
         if output_dir is None:
             exit("When running in Docker spawn mode we need an output dir.")
-        docker = Docker(output_dir)
-        docker.run()
-        http_port = docker.http_port
-        bolt_port = docker.bolt_port
+        container = Docker(output_dir)
+        container.run()
+        http_port = container.http_port
+        bolt_port = container.bolt_port
     else:
         http_port = 7474
         bolt_port = 7687
 
-    neo_db = NeoDb(host=os.environ.get("DATABASE_URL", "localhost"), password="",
+    neo_db = NeoDb(host=HOST, password="",
                    use_bolt=True, bolt_port=bolt_port, http_port=http_port)
     start = time.time()
-    process_vcf_files(neo_db, vcf_dir=vcf_dir,
-                      owner=owner, history_id=history_id)
-    if d:
-        docker.stop()
+    antibiotic = '\t'.join(antibiotic)
+    process_vcf_files(neo_db, vcf_dir=vcf_dir, phenotype=phenotype,
+                      antibiotic=antibiotic, owner=owner,
+                      history_id=history_id)
+    if docker:
+        container.stop()
     end = time.time()
     sys.stdout.write(f"\nDone in {end - start} ms.\n")
 

vcf2neo/db.py

@@ -120,12 +120,13 @@ def create_variant_site_nodes(self, record, known_sites,
             self.graph.push(v_set)
         return known_sites
 
-    def create_call_set_nodes(self, set_name, v_set):
+    def create_call_set_nodes(self, set_name, phenotype, antibiotic, v_set):
         """
         Create CallSet Nodes
         :return:
         """
-        c_set = CallSet(name=set_name)
+        c_set = CallSet(name=set_name, phenotype=phenotype,
+                        antibiotic=antibiotic)
         c_set.belongs_to_vset.add(v_set)
         self.graph.create(c_set)
         return c_set

vcf2neo/vcfproc.py

@@ -10,7 +10,8 @@
 import vcf
 
 
-def process_vcf_files(db, vcf_dir, owner=None, history_id=None):
+def process_vcf_files(db, vcf_dir, phenotype, antibiotic, owner=None,
+                      history_id=None):
     owner = owner if owner else getpass.getuser()
     known_sites = dict()
     if os.path.isdir(vcf_dir):
@@ -31,7 +32,8 @@ def process_vcf_files(db, vcf_dir, owner=None, history_id=None):
                             c_set_name = os.path.basename(
                                 os.path.abspath(_file))
                             c_set = db.create_call_set_nodes(
-                                set_name=c_set_name, v_set=v_set
+                                set_name=c_set_name, v_set=v_set,
+                                phenotype=phenotype, antibiotic=antibiotic
                             )
                             known_sites = get_variant_sites(
                                 db, known_sites, vcf_reader, v_set=v_set,