pyprocessors-nameparser 0.5.3__tar.gz

pyprocessors_nameparser-0.5.3/.dockerignore

@@ -0,0 +1,9 @@
+# ignore .git and .cache folders
+.git
+.cache
+.pytest_cache
+models
+tests
+log
+**/__pycache__
+*.pyc

pyprocessors_nameparser-0.5.3/.gitignore

@@ -0,0 +1,20 @@
+/.idea/
+/.vscode/
+/dist/
+dist
+.git
+.cache
+**/.pytest_cache
+models
+log
+**/__pycache__/
+*.pyc
+**/test-reports
+**/.passwd-pypi
+**/.env
+**/results.xml
+.groovylintrc.json
+.emailNotif
+
+# Unit test / coverage reports
+.tox/

pyprocessors_nameparser-0.5.3/Dockerfile

@@ -0,0 +1,19 @@
+FROM python:3.8-slim-buster
+# Install prerequisites
+RUN apt-get update -y && \
+    apt-get install -y \
+    patch \
+    gcc && \
+    apt-get install -y --no-install-recommends \
+    g++ \
+    git && \
+# Final upgrade + clean
+    apt-get update -y && \
+    apt-get clean all -y
+
+# Enable Installing packages as root
+ENV FLIT_ROOT_INSTALL=1
+
+# Add pyproject.toml + README.md for flit install
+ADD pyproject.toml pyproject.toml
+ADD README.md README.md

pyprocessors_nameparser-0.5.3/Jenkinsfile

@@ -0,0 +1,408 @@
+pipeline {
+  environment {
+    PATH_HOME = '/home/jenkins'
+    TEST_REPORT_DIR = '/root/test-reports'
+    PYTHONPYCACHEPREFIX = '/tmp/.pytest_cache'
+    PYTHONDONTWRITEBYTECODE = '1'
+    JENKINS_UIDGID = '1004:1004'
+
+    MAJOR_VERSION = '0'
+    MINOR_VERSION = '5'
+  }
+
+  agent none
+
+  triggers {
+    upstream(upstreamProjects: 'pymultirole_plugins/' + BRANCH_NAME.replaceAll('/', '%2F'),\
+                                threshold: hudson.model.Result.SUCCESS)
+  }
+
+  stages {
+    stage('Catch build termination') {
+      agent {
+        node {
+          label 'built-in'
+          customWorkspace "${PATH_HOME}/${JOB_NAME}"
+        }
+      }
+      stages {
+        stage('Analyse build cause') {
+          steps {
+            script {
+              analyseBuildCause()
+            }
+          }
+        }
+      }
+    }
+
+    stage('Generate new version') {
+      when {
+        environment name: 'SKIP_JOB', value: '0'
+      }
+
+      agent {
+        node {
+          label 'built-in'
+          customWorkspace "${PATH_HOME}/${JOB_NAME}"
+        }
+      }
+
+      stages {
+        stage('Add credentials') {
+          steps {
+            script {
+              // Add password file for flit publishing
+              sh "cp ${PATH_HOME}/.passwd-pypi .env"
+            }
+          }
+        }
+
+        stage('Commit new version') {
+          steps {
+            script {
+              println("attempt to publish ${JOB_NAME} with version: ${MAJOR_VERSION}.${MINOR_VERSION}.${BUILD_ID}")
+
+              // push updates of file __init__.py
+              withCredentials([gitUsernamePassword(credentialsId: 'bitbucket-user', gitToolName: 'git-tool')]) {
+                sh 'git pull'
+                sh "echo '\"\"\"Processor based on Nameparser\"\"\"' > pyprocessors_nameparser/__init__.py"
+                sh "echo '__version__ = \"${MAJOR_VERSION}.${MINOR_VERSION}.${BUILD_ID}\"' >> pyprocessors_nameparser/__init__.py"
+                sh 'git commit pyprocessors_nameparser/__init__.py -m "[Jenkins CI] Commit on version files" || echo "No changes to commit"'
+                sh 'git push'
+              }
+            }
+          }
+        }
+      }
+    }
+
+    stage('Build, test and publish') {
+      when {
+        beforeAgent true
+        environment name: 'SKIP_JOB', value: '0'
+      }
+
+      agent {
+        // dockerfile agent
+        // Mounted volume for Junit reports
+        //   - docker: /root/test-reports
+        //   - host  : /tmp/_${JOB_NAME}/test-reports
+        dockerfile {
+          label 'built-in'
+          customWorkspace "${PATH_HOME}/${JOB_NAME}"
+          filename 'Dockerfile'
+          args "-u root --privileged -v /tmp/_${JOB_NAME}/test-reports:${TEST_REPORT_DIR}"
+        }
+      }
+
+      stages {
+        stage('Install flit & flake8') {
+          steps {
+            // remove any previous tox env
+            sh 'rm -rf .tox'
+            sh 'python -m pip install pip==22.0.3'
+            sh 'pip install --no-cache-dir flit==3.2.0 flake8==3.9.2 flakehell tox'
+            sh 'flit install'
+          }
+        }
+
+        stage('Test & lint python code') {
+          steps {
+            // remove any previous results.xml file
+            sh "rm -f ${TEST_REPORT_DIR}/results.xml"
+            sh 'tox'
+          }
+        }
+
+        stage('Publish on PyPI') {
+          environment {
+            FLIT_USERNAME = getUserName '.env'
+            FLIT_PASSWORD = getUserPass '.env'
+          }
+          steps {
+            // remove any previous folder dist
+            sh 'rm -rf dist'
+            // create (as root) folder dist
+            sh 'mkdir dist'
+            // pull recent updates of file __init__.py
+            withCredentials([gitUsernamePassword(credentialsId: 'bitbucket-user', gitToolName: 'git-tool')]) {
+              sh 'git config --global pull.rebase false'
+              sh "git config --global --add safe.directory ${WORKSPACE}"
+              sh 'git pull'
+            }
+            // put back owner of .git folder
+            sh "chown -R ${JENKINS_UIDGID} ${WORKSPACE}/.git"
+            // put back owner of pulled file
+            sh "chown ${JENKINS_UIDGID} pyprocessors_nameparser/__init__.py"
+            // get git status
+            sh 'git status'
+            // publish on PyPI
+            sh '''
+               export COMMIT_VERSION=$( cat pyprocessors_nameparser/__init__.py|grep version|cut -d '"' -f2|tr -s '[:blank:]' )
+               export BUILD_VERSION="${MAJOR_VERSION}"."${MINOR_VERSION}"."${BUILD_ID}"
+               if [ "${COMMIT_VERSION}" = "${BUILD_VERSION}" ] ; then flit publish ; fi
+               '''
+            // remove current folder dist
+            sh 'rm -rf dist'
+            // remove current folder .hypothesis
+            sh 'rm -rf .hypothesis'
+            // remove current folder .tox
+            sh 'rm -rf .tox'
+          }
+        }
+      }
+    }
+  }
+
+  post {
+    // only triggered when blue or green sign
+    success {
+      // node is specified here to get an agent
+      node('built-in') {
+        // keep using customWorkspace to store Junit report
+        ws("${PATH_HOME}/${JOB_NAME}") {
+          script {
+            try {
+              sh 'rm -f results.xml'
+              sh "cp /tmp/_${JOB_NAME}/test-reports/results.xml results.xml"
+            } catch (Exception e) {
+              println 'Exception occurred: ' + e.toString()
+            }
+            try {
+              junit 'results.xml'
+            } catch (Exception e) {
+              println 'Exception occurred: ' + e.toString()
+            }
+            if (sendEmailNotif("${PATH_HOME}/${JOB_NAME}", "${BUILD_NUMBER}")) {
+              println 'sending Success Build notification'
+              CUSTOM_SUBJECT = '[CI - Jenkinzz SUCCESS] ' + CUSTOM_SUBJECT
+              emailext(
+                  mimeType: 'text/html',
+                  subject: CUSTOM_SUBJECT,
+                  body: '${DEFAULT_CONTENT}',
+                  replyTo: '${DEFAULT_REPLYTO}',
+                  to: '${ADMIN_RECIPIENTS}' + ';' + CUSTOM_RECIPIENTS
+              )
+              switchEmailNotif(false, BUILD_NUMBER)
+            } else {
+              println 'preventing Success Build notification'
+            }
+          }
+        }
+      }
+    }
+    // triggered when red sign
+    failure {
+      // node is specified here to get an agent
+      node('built-in') {
+        // keep using customWorkspace to store Junit report
+        ws("${PATH_HOME}/${JOB_NAME}") {
+          script {
+            try {
+              sh 'rm -f results.xml'
+              sh "cp /tmp/_${JOB_NAME}/test-reports/results.xml results.xml"
+            } catch (Exception e) {
+              println 'Exception occurred: ' + e.toString()
+            }
+            try {
+              junit 'results.xml'
+            } catch (Exception e) {
+              println 'Exception occurred: ' + e.toString()
+            }
+            println 'sending Failure Build notification'
+            CUSTOM_SUBJECT = '[CI - Jenkinzz FAILURE] ' + CUSTOM_SUBJECT
+            emailext(
+                mimeType: 'text/html',
+                subject: CUSTOM_SUBJECT,
+                body: '${DEFAULT_CONTENT}',
+                replyTo: '${DEFAULT_REPLYTO}',
+                to: '${ADMIN_RECIPIENTS}' + ';' + CUSTOM_RECIPIENTS
+            )
+          }
+        }
+      }
+    }
+    // triggered when black sign
+    aborted {
+      println 'post-declarative message: abort job'
+    }
+    // trigger every-works
+    //always {
+    //}
+  }
+}
+
+// return FLIT_USERNAME from given file
+def getUserName(path) {
+  USERNAME = sh(
+                 script: "grep FLIT_USERNAME ${path}|cut -d '=' -f2",
+                 returnStdout: true
+               ).trim()
+  return USERNAME
+}
+
+// return FLIT_PASSWORD from given file
+def getUserPass(path) {
+  USERPASS = sh(
+                 script: "grep FLIT_PASSWORD ${path}|cut -d '=' -f2",
+                 returnStdout: true
+               ).trim()
+  return USERPASS
+}
+
+// create/remove emailNotif file to trigger email notification
+def switchEmailNotif(toggle, build) {
+  if (toggle) {
+    sh 'echo ' + build + ' > .emailNotif'
+  } else {
+    if (build == BUILD_NUMBER) {
+      sh 'rm -f .emailNotif'
+    }
+  }
+}
+
+// return true if emailNotif file present
+boolean sendEmailNotif(path, build) {
+  emailNotif = sh(
+                 script: "find ${path} -name '.emailNotif'|wc -l",
+                 returnStdout: true
+               ).trim()
+  emailContent = ''
+  if (emailNotif == '1') {
+    emailContent = sh(
+                     script: "cat ${path}/.emailNotif",
+                     returnStdout: true
+                   ).trim()
+  }
+  return (emailContent == build)
+}
+
+def analyseBuildCause() {
+  upstreamProjects = ['pyimporters_plugins']
+  boolean upstreamRunning = false
+  String jobName
+  // iterate over upstreamProjects
+  for (upstream_project in upstreamProjects) {
+    Jenkins.instance.getItemByFullName(upstream_project).items.each { repository ->
+      boolean isRunning = false
+      //repository.parent.name: project
+      //repository.name: branch
+      if ( repository.name == BRANCH_NAME ) {
+        // iterate over all jobs of current repository
+        repository.allJobs.each { job ->
+          // iterate over all builds of current job
+          job.builds.each { build ->
+            // determine if a build is running or not
+            if ( build.result == (null) ) {
+              jobName = build.parent.parent.name
+              isRunning = true
+            }
+          }
+          if ( isRunning ) {
+            upstreamRunning = true
+          }
+        }
+      }
+    }
+  }
+
+  // Catch if build has been triggered by CI Commit
+  // returnStatus = true  when string not found -> Team commit
+  // returnStatus = false when string is found  -> CI commit
+  boolean lastCommitIsTeam = sh(
+    script: 'git log -1 | grep "\\[Jenkins CI\\]"',
+    returnStatus: true
+  )
+
+  // Skip build when upstream detected
+  if (upstreamRunning) {
+    println 'Skipping build because upstream job detected (' + jobName + ')'
+    env.SKIP_JOB = '1'
+    switchEmailNotif(false, 0)
+    currentBuild.result = 'NOT_BUILT'
+  }
+
+  // Catch if build has been triggered by User
+  boolean isStartedByUser = currentBuild.rawBuild.getCause(hudson.model.Cause$UserIdCause) != null
+  if (isStartedByUser && !upstreamRunning) {
+    env.SKIP_JOB = '0'
+    env.CUSTOM_SUBJECT = JOB_NAME + ' - Manual Build #' + BUILD_NUMBER
+    env.CUSTOM_RECIPIENTS = emailextrecipients([[$class: 'RequesterRecipientProvider']])
+    switchEmailNotif(true, BUILD_NUMBER)
+    println 'Job started by User, proceeding'
+  }
+
+  // Catch if build has been triggered by Upstream
+  boolean isStartedByUpstream = currentBuild.rawBuild.getCause(hudson.model.Cause$UpstreamCause) != null
+  if (isStartedByUpstream && !upstreamRunning) {
+    int changeSetCount = 0
+    int ciSkipCount = 0
+    String upstreamFullJobName = ''
+    for (Run upstreamBuild : currentBuild.upstreamBuilds) {
+      upstreamFullJobName   = upstreamBuild.rawBuild.fullDisplayName
+      if (upstreamBuild.changeSets != null) {
+        def changeLogSets = upstreamBuild.changeSets
+        for (int i = 0; i < changeLogSets.size(); i++) {
+          changeSetCount++
+          def entries = changeLogSets[i].items
+          for (int j = 0; j < entries.length; j++) {
+            def entry = entries[j]
+            if (entry.msg.contains('[Jenkins CI]')) {
+              ciSkipCount++
+            }
+          }
+        }
+      }
+    }
+    if (changeSetCount > 0 && changeSetCount == ciSkipCount) {
+      env.SKIP_JOB = '1'
+      switchEmailNotif(false, 0)
+      println 'Job started by Upstream [' + upstreamFullJobName + '], with CI commit, skipping'
+      currentBuild.result = 'NOT_BUILT'
+    } else {
+      env.SKIP_JOB = '0'
+      env.CUSTOM_SUBJECT = JOB_NAME + ' - Upstream Build #' + BUILD_NUMBER
+      env.CUSTOM_RECIPIENTS = emailextrecipients([[$class:'UpstreamComitterRecipientProvider']])
+      switchEmailNotif(true, BUILD_NUMBER)
+      println 'Job started by Upstream [' + upstreamFullJobName + '], proceeding'
+    }
+  }
+
+  // Catch if build has been triggered by User Commit
+  boolean isStartedByCommit = currentBuild.rawBuild.getCause(jenkins.branch.BranchEventCause) != null
+  if (isStartedByCommit && lastCommitIsTeam && !upstreamRunning) {
+    env.SKIP_JOB = '0'
+    env.CUSTOM_SUBJECT = JOB_NAME + ' - SCM Build #' + BUILD_NUMBER
+    env.CUSTOM_RECIPIENTS = emailextrecipients([[$class: 'DevelopersRecipientProvider'], [$class:'CulpritsRecipientProvider']])
+    switchEmailNotif(true, BUILD_NUMBER)
+    println 'Job started by User Commit, proceeding'
+  }
+
+  // Catch if build has been triggered by cron
+  boolean isStartedByCron = currentBuild.rawBuild.getCause(hudson.triggers.TimerTrigger$TimerTriggerCause) != null
+  if (isStartedByCron && lastCommitIsTeam && !upstreamRunning) {
+    env.SKIP_JOB = '0'
+    env.CUSTOM_SUBJECT = JOB_NAME + ' - CRON Build #' + BUILD_NUMBER
+    env.CUSTOM_RECIPIENTS = emailextrecipients([[$class: 'DevelopersRecipientProvider'], [$class:'CulpritsRecipientProvider']])
+    switchEmailNotif(true, BUILD_NUMBER)
+    println 'Job started by Cron, proceeding'
+  }
+
+  // Catch if build has been triggered by branch discovery
+  boolean isStartedByBranchDiscovery = currentBuild.rawBuild.getCause(jenkins.branch.BranchIndexingCause) != null
+  if (isStartedByBranchDiscovery && lastCommitIsTeam && !upstreamRunning) {
+    env.SKIP_JOB = '0'
+    env.CUSTOM_SUBJECT = JOB_NAME + ' - BranchDiscovery Build #' + BUILD_NUMBER
+    env.CUSTOM_RECIPIENTS = emailextrecipients([[$class: 'DevelopersRecipientProvider'], [$class:'CulpritsRecipientProvider']])
+    switchEmailNotif(true, BUILD_NUMBER)
+    println 'Job started by Branch Discovery, proceeding'
+  }
+
+  if (!lastCommitIsTeam && !upstreamRunning && !isStartedByUser && !isStartedByUpstream) {
+    println 'Skipping build because last commit has been done by CI'
+    env.SKIP_JOB = '1'
+    switchEmailNotif(false, 0)
+    //currentBuild.result = 'NOT_BUILT'
+  }
+}

pyprocessors_nameparser-0.5.3/PKG-INFO

@@ -0,0 +1,66 @@
+Metadata-Version: 2.3
+Name: pyprocessors-nameparser
+Version: 0.5.3
+Summary: Processor based on Nameparser
+Home-page: https://kairntech.com/
+Author: Olivier Terrier
+Author-email: olivier.terrier@kairntech.com
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+Classifier: Intended Audience :: Information Technology
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: System Administrators
+Classifier: Operating System :: OS Independent
+Classifier: Topic :: Software Development :: Libraries :: Application Frameworks
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Topic :: Software Development :: Libraries
+Classifier: Topic :: Software Development
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Development Status :: 4 - Beta
+Classifier: Programming Language :: Python :: 3.8
+Requires-Dist: pymultirole_plugins>=0.5.0,<0.6.0
+Requires-Dist: log-with-context
+Requires-Dist: nameparser
+Requires-Dist: flit ; extra == "dev"
+Requires-Dist: pre-commit ; extra == "dev"
+Requires-Dist: bump2version ; extra == "dev"
+Requires-Dist: sphinx ; extra == "docs"
+Requires-Dist: sphinx-rtd-theme ; extra == "docs"
+Requires-Dist: m2r2 ; extra == "docs"
+Requires-Dist: sphinxcontrib.apidoc ; extra == "docs"
+Requires-Dist: jupyter_sphinx ; extra == "docs"
+Requires-Dist: pytest ; extra == "test"
+Requires-Dist: pytest-cov ; extra == "test"
+Requires-Dist: pytest-flake8 ; extra == "test"
+Requires-Dist: pytest-black ; extra == "test"
+Requires-Dist: flake8==3.9.2 ; extra == "test"
+Requires-Dist: tox ; extra == "test"
+Requires-Dist: dirty-equals ; extra == "test"
+Provides-Extra: dev
+Provides-Extra: docs
+Provides-Extra: test
+
+## Requirements
+
+- Python 3.8+
+- Flit to put Python packages and modules on PyPI
+- Pydantic for the data parts.
+
+## Installation
+```
+pip install flit
+pip install pymultirole-plugins
+```
+
+## Publish the Python Package to PyPI
+- Increment the version of your package in the `__init__.py` file:
+```
+"""An amazing package!"""
+
+__version__ = 'x.y.z'
+```
+- Publish
+```
+flit publish
+```
+

pyprocessors_nameparser-0.5.3/README.md

@@ -0,0 +1,23 @@
+## Requirements
+
+- Python 3.8+
+- Flit to put Python packages and modules on PyPI
+- Pydantic for the data parts.
+
+## Installation
+```
+pip install flit
+pip install pymultirole-plugins
+```
+
+## Publish the Python Package to PyPI
+- Increment the version of your package in the `__init__.py` file:
+```
+"""An amazing package!"""
+
+__version__ = 'x.y.z'
+```
+- Publish
+```
+flit publish
+```

pyprocessors_nameparser-0.5.3/bumpversion.py

@@ -0,0 +1,41 @@
+import re
+import sys
+from pathlib import Path
+
+
+def main(argv):
+    part = argv[0].lower() if len(argv) > 0 else "minor"
+    Jenkinsfile = Path("./Jenkinsfile")
+    pyprojectfile = Path("./pyproject.toml")
+    with Jenkinsfile.open("r", encoding="utf-8") as fin:
+        data = fin.read()
+        for line in data.split(fin.newlines):
+            if "MAJOR_VERSION" in line:
+                result = re.search(r'MAJOR_VERSION\s*=\s*"([0-9]+)"', line)
+                if result:
+                    major = int(result.group(1))
+                    if part == "major":
+                        to_replace = result.group(0)
+                        by_replace = f'MAJOR_VERSION = "{major + 1}"'
+            if "MINOR_VERSION" in line:
+                result = re.search(r'MINOR_VERSION\s*=\s*"([0-9]+)"', line)
+                if result:
+                    minor = int(result.group(1))
+                    if part == "minor":
+                        to_replace = result.group(0)
+                        by_replace = f'MINOR_VERSION = "{minor + 1}"'
+    data = data.replace(to_replace, by_replace)
+    with Jenkinsfile.open("wt") as fout:
+        fout.write(data)
+
+    depend_string = f"{major}.{minor}.0,<{major}.{minor + 1}.0"
+    new_depend_string = f"{major}.{minor + 1}.0,<{major}.{minor + 2}.0"
+    with pyprojectfile.open("r", encoding="utf-8") as fin:
+        data = fin.read()
+    data = data.replace(depend_string, new_depend_string)
+    with pyprojectfile.open("wt") as fout:
+        fout.write(data)
+
+
+if __name__ == "__main__":
+    main(sys.argv[1:])

pyprocessors_nameparser-0.5.3/pyprocessors_nameparser/__init__.py

@@ -0,0 +1,2 @@
+"""Processor based on Nameparser"""
+__version__ = "0.5.3"

pyprocessors_nameparser-0.5.3/pyprocessors_nameparser/name_parser.py

@@ -0,0 +1,40 @@
+from logging import Logger
+from typing import List, Type, cast
+from log_with_context import add_logging_context
+from nameparser import HumanName
+from pydantic import Field, BaseModel
+from pymultirole_plugins.v1.processor import ProcessorParameters, ProcessorBase
+from pymultirole_plugins.v1.schema import Document
+
+logger = Logger("pymultirole")
+
+
+class NameParserParameters(ProcessorParameters):
+    name_labels: List[str] = Field(None, description="List of labels to analyze", extra="label")
+
+
+class NameParserProcessor(ProcessorBase):
+    __doc__ = """NameParser based on [Nameparser](https://github.com/derek73/python-nameparser)."""
+
+    def process(
+            self, documents: List[Document], parameters: ProcessorParameters
+    ) -> List[Document]:
+        params: NameParserParameters = cast(NameParserParameters, parameters)
+        try:
+            for document in documents:
+                with add_logging_context(docid=document.identifier):
+                    if document.annotations:
+                        for a in document.annotations:
+                            if a.labelName in params.name_labels:
+                                atext = a.text or document.text[a.start:a.end]
+                                name = HumanName(atext)
+                                props = a.properties or {}
+                                props.update(name.as_dict())
+                                a.properties = props
+        except BaseException as err:
+            raise err
+        return documents
+
+    @classmethod
+    def get_model(cls) -> Type[BaseModel]:
+        return NameParserParameters