added support for ES 7 (fixes #155) (#161)

.ci/seed_es_on_travis.sh

@@ -17,104 +17,109 @@ TEST_DATA_DIR=$(pwd)/test_data
 LEGACY_MAPPING_FILE="${TEST_DATA_DIR}/legacy_shakespeare_mapping.json"
 ES5_MAPPING_FILE="${TEST_DATA_DIR}/es5_shakespeare_mapping.json"
 ES6_MAPPING_FILE="${TEST_DATA_DIR}/es6_shakespeare_mapping.json"
+ES7_MAPPING_FILE="${TEST_DATA_DIR}/es7_shakespeare_mapping.json"
 
+SAMPLE_DATA_FILE=${TEST_DATA_DIR}/sample.json
+ES7_SAMPLE_DATA_FILE=${TEST_DATA_DIR}/sample_es7.json
 
 case "$ES_VERSION" in
     "") ;;
 
     "1.0.0")
-      export ES_VERSION=1.0.0;
       export MAPPING_FILE=${LEGACY_MAPPING_FILE};
       export ES_BINARY_URL="${ES1_ARCHIVE}/elasticsearch-$ES_VERSION.deb"
       ;;
 
     "1.4.4")
-      export ES_VERSION=1.4.4;
       export MAPPING_FILE=${LEGACY_MAPPING_FILE};
       export ES_BINARY_URL="${ES1_ARCHIVE}/elasticsearch-$ES_VERSION.deb"
       ;;
 
     "1.7.2")
-      export ES_VERSION=1.7.2;
       export MAPPING_FILE=${LEGACY_MAPPING_FILE};
       export ES_BINARY_URL="${ES1_ARCHIVE}/elasticsearch-$ES_VERSION.deb"
       ;;
 
     "2.0.2")
-      export ES_VERSION=2.0.2;
       export MAPPING_FILE=${LEGACY_MAPPING_FILE};
       export ES_BINARY_URL="${ES2_ARCHIVE}/$ES_VERSION/elasticsearch-$ES_VERSION.deb"
       ;;
 
     "2.1.2")
-      export ES_VERSION=2.1.2;
       export MAPPING_FILE=${LEGACY_MAPPING_FILE};
       export ES_BINARY_URL="${ES2_ARCHIVE}/$ES_VERSION/elasticsearch-$ES_VERSION.deb"
       ;;
 
     "2.2.2")
-      export ES_VERSION=2.2.2;
       export MAPPING_FILE=${LEGACY_MAPPING_FILE};
       export ES_BINARY_URL="${ES2_ARCHIVE}/$ES_VERSION/elasticsearch-$ES_VERSION.deb"
       ;;
 
     "2.3.5")
-      export ES_VERSION=2.3.5;
       export MAPPING_FILE=${LEGACY_MAPPING_FILE};
       export ES_BINARY_URL="${ES2_ARCHIVE}/$ES_VERSION/elasticsearch-$ES_VERSION.deb"
       ;;
 
     "5.0.2")
-      export ES_VERSION=5.0.2;
       export MAPPING_FILE=${ES5_MAPPING_FILE};
       export ES_BINARY_URL="${ES5PLUS_ARCHIVE}/elasticsearch-$ES_VERSION.deb"
       ;;
 
     "5.3.3")
-      export ES_VERSION=5.3.3;
       export MAPPING_FILE=${ES5_MAPPING_FILE};
       export ES_BINARY_URL="${ES5PLUS_ARCHIVE}/elasticsearch-$ES_VERSION.deb"
       ;;
-    
+
     "5.4.3")
-      export ES_VERSION=5.4.3;
       export MAPPING_FILE=${ES5_MAPPING_FILE};
       export ES_BINARY_URL="${ES5PLUS_ARCHIVE}/elasticsearch-$ES_VERSION.deb"
       ;;
 
     "5.6.9")
-      export ES_VERSION=5.6.9;
       export MAPPING_FILE=${ES5_MAPPING_FILE};
       export ES_BINARY_URL="${ES5PLUS_ARCHIVE}/elasticsearch-$ES_VERSION.deb"
       ;;
-    
+
     "6.0.1")
-      export ES_VERSION=6.0.1;
       export MAPPING_FILE=${ES6_MAPPING_FILE};
       export ES_BINARY_URL="${ES5PLUS_ARCHIVE}/elasticsearch-$ES_VERSION.deb"
       ;;
-    
+
     "6.1.4")
       export ES_VERSION=6.1.4;
       export MAPPING_FILE=${ES6_MAPPING_FILE};
       export ES_BINARY_URL="${ES5PLUS_ARCHIVE}/elasticsearch-$ES_VERSION.deb"
       ;;
-      
+
     "6.2.4")
-      export ES_VERSION=6.2.4;
       export MAPPING_FILE=${ES6_MAPPING_FILE};
       export ES_BINARY_URL="${ES5PLUS_ARCHIVE}/elasticsearch-$ES_VERSION.deb"
       ;;
+
+    "7.3.1")
+      export MAPPING_FILE=${ES7_MAPPING_FILE};
+      export ES_BINARY_URL="${ES5PLUS_ARCHIVE}/elasticsearch-$ES_VERSION-amd64.deb"
+      # overwrite SAMPLE_DATA_FILE to use the ES7-compliant data
+      export SAMPLE_DATA_FILE="${ES7_SAMPLE_DATA_FILE}"
+      ;;
    esac
 
 # pull the binary
-curl -O ${ES_BINARY_URL}
+curl ${ES_BINARY_URL} \
+    --output elasticsearch.deb
 
 # start the service and wait a bit
 sudo dpkg \
     -i \
     --force-confnew \
-    elasticsearch-$ES_VERSION.deb
+    elasticsearch.deb
+
+# deal with permissions
+# reference: https://discuss.elastic.co/t/permission-denied-starting-elasticsearch-7-0/179336
+sudo chown -R \
+    elasticsearch:elasticsearch \
+    /etc/default/elasticsearch
+
 sudo service elasticsearch start
 sleep ${SLEEP_TIL_STARTUP_SECONDS}
 sudo service elasticsearch status
@@ -135,7 +140,7 @@ curl -X PUT \
     -H 'Content-Type:application/json' \
     -d @shakespeare_mapping.json
 
-mv test_data/sample.json sample.json
+mv ${SAMPLE_DATA_FILE} sample.json
 
 curl -X POST \
     "${ES_HOST}/shakespeare/_bulk" \

.ci/setup.sh

@@ -11,7 +11,7 @@ export JAVA_APT_PKG="oracle-java8-set-default"
 # install these testing packages we need
 if [[ "$TASK" == "rpkg" ]];
 then
-  Rscript -e "install.packages(c('devtools', 'knitr', 'testthat', 'rmarkdown'), repos = 'http://cran.rstudio.com')"
+  Rscript -e "install.packages(c('data.table', 'devtools', 'futile.logger', 'knitr', 'testthat', 'rmarkdown', 'uuid'), repos = 'http://cran.rstudio.com')"
   cp test_data/* r-pkg/inst/testdata/
 fi
 

.travis.yml

@@ -30,6 +30,7 @@ script:
 # - ES_VERSION=6.0.1
 # - ES_VERSION=6.1.4
 # - ES_VERSION=6.2.4
+# - ES_VERSION=7.3.1
 matrix:
   include:
     ############
@@ -119,6 +120,12 @@ matrix:
       env:
         - ES_VERSION=6.2.4
         - TASK=rpkg
+    - language: r
+      warnings_are_errors: true
+      cache: packages
+      env:
+        - ES_VERSION=7.3.1
+        - TASK=rpkg
       after_success:
         - .ci/report_to_covr.sh
     #################
@@ -194,3 +201,8 @@ matrix:
       env:
         - ES_VERSION=6.2.4
         - TASK=pypkg
+    - language: python
+      python: 3.5
+      env:
+        - ES_VERSION=7.3.1
+        - TASK=pypkg

NEWS.md

@@ -1,3 +1,10 @@
+# uptasticsearch development version
+
+## Features
+
+### Added support for ES7.x
+- [#161](https://github.com/uptake/uptasticsearch/pull/161) Added support for ES7.x. The biggest changes between that major version and 6.x were the removal of `_all` as a way to reference all indices, changing the response format of `hits.total` into an object like `{"hits": {"total": 50}}`, and restricting all indices to have a single type of document. More details can be found at https://www.elastic.co/guide/en/elasticsearch/reference/current/breaking-changes-7.0.html.
+
 # uptasticsearch 0.3.1
 
 ## Bugfixes

py-pkg/setup.py

@@ -8,7 +8,7 @@
 import subprocess
 import datetime
 
-CURRENT_VERSION = "0.1.0"
+CURRENT_VERSION = "0.2.0"
 
 #########################
 

py-pkg/tests/test_clients.py

@@ -8,21 +8,26 @@
 from uptasticsearch.clients import Uptasticsearch2
 from uptasticsearch.clients import Uptasticsearch5
 from uptasticsearch.clients import Uptasticsearch6
+from uptasticsearch.clients import Uptasticsearch7
 from uptasticsearch.clients import uptasticsearch_factory
 
 from uptasticsearch.fetch_all import es_search
 
 
 class TestEsSearch(object):
     """
-    es_search should work an return a pandas DataFrame
+    es_search should work and return a pandas DataFrame
     """
     host = "http://127.0.0.1:9200"
+
     def test_rectangle(self):
-        assert isinstance(es_search(self.host,
-                                    "shakespeare",
-                                    query_body=json.dumps({}),
-                                    size=10000,
-                                    max_hits=10,
-                                    scroll="1m"),
-                          pd.DataFrame)
+        result = es_search(
+            self.host,
+            "shakespeare",
+            query_body=json.dumps({}),
+            size=10000,
+            max_hits=10,
+            scroll="1m"
+        )
+        print(result)
+        assert isinstance(result, pd.DataFrame)

py-pkg/uptasticsearch/clients.py

@@ -33,6 +33,14 @@ def __init__(self, url, http_client=HttpClient()):
         self.url = _format_es_url(url)
         self.client = http_client
 
+    def _get_total_hits(self, response_json):
+        """
+        Given a dictionary representing the content of a
+        respose to a  ``POST /_search`` request, return the total
+        number of docs matching the query
+        """
+        return response_json['hits']['total']
+
     def search(self, body, index="", doc_type="", scroll_context_timer="1m", page_size=10000,  max_hits=None):
         """Execute a Search Query on the Elasticsearch Cluster
 
@@ -60,7 +68,7 @@ def search(self, body, index="", doc_type="", scroll_context_timer="1m", page_si
                                     headers={'Content-Type': 'application/json'})
 
         page = response.json()
-        total_hits = page['hits']['total']
+        total_hits = self._get_total_hits(page)
         total_hits = min(total_hits, max_hits) if max_hits is not None else total_hits
 
         results = [d['_source'] for d in page['hits']['hits']]
@@ -118,13 +126,26 @@ def _make_scroll_request(self, scroll_context_timer, scroll_id):
                                                  "scroll_id": scroll_id}),
                                 headers={'Content-Type': 'application/json'})
 
+class Uptasticsearch7(Uptasticsearch6):
+
+    def _get_total_hits(self, response_json):
+        """
+        Given a dictionary representing the content of a
+        respose to a  ``POST /_search`` request, return the total
+        number of docs matching the query
+        """
+        return response_json['hits']['total']['value']
+
 
 def uptasticsearch_factory(url, retries=5, backoff_factor=0.1):
     http_client = HttpClient(retries=retries, backoff_factor=backoff_factor)
     es_url = _format_es_url(url)
     cluster_version = http_client.get(es_url).json()['version']['number'].split('.')[0]
 
-    return {"1": Uptasticsearch1,
-            "2": Uptasticsearch2,
-            "5": Uptasticsearch5,
-            "6": Uptasticsearch6}[cluster_version](es_url, http_client)
+    return {
+        "1": Uptasticsearch1,
+        "2": Uptasticsearch2,
+        "5": Uptasticsearch5,
+        "6": Uptasticsearch6,
+        "7": Uptasticsearch7
+    }[cluster_version](es_url, http_client)

r-pkg/DESCRIPTION

@@ -1,7 +1,7 @@
 Package: uptasticsearch
 Type: Package
 Title: Get Data Frame Representations of 'Elasticsearch' Results
-Version: 0.3.1
+Version: 0.4.1
 Authors@R: c(
     person("James", "Lamb", email = "jaylamb20@gmail.com", role = c("aut", "cre")),
     person("Nick", "Paras", email = "nick.paras@uptake.com", role = c("aut")),

r-pkg/R/es_search.R

@@ -296,7 +296,13 @@ es_search <- function(es_host
 
     # Parse to JSON to get total number of documents matching the query
     firstResult <- jsonlite::fromJSON(firstResultJSON, simplifyVector = FALSE)
-    hits_to_pull <- min(firstResult[["hits"]][["total"]], max_hits)
+
+    major_version <- .get_es_version(es_host)
+    if (as.integer(major_version) > 6){
+      hits_to_pull <- min(firstResult[["hits"]][["total"]][["value"]], max_hits)
+    } else {
+      hits_to_pull <- min(firstResult[["hits"]][["total"]], max_hits)
+    }
 
     # If we got everything possible, just return here
     hits_pulled <- length(firstResult[["hits"]][["hits"]])