Remove empty lines in HTTP examples, reformat some of them

qdrant-landing/content/articles/memory-consumption.md

@@ -173,8 +173,7 @@ In the second experiment, we tested how well our system performs when **vectors
 Create collection with:
 
 ```http
-PUT /collections/benchmark 
-
+PUT /collections/benchmark
 {
   ...
   "optimizers_config": {
@@ -222,7 +221,6 @@ Create collection with:
 
 ```http
 PUT /collections/benchmark 
-
 {
   ...
   "hnsw_config": {

qdrant-landing/content/articles/new-recommendation-api.md

@@ -35,9 +35,8 @@ combine them within a single request. That makes the new implementation backward
 Qdrant instance without any changes in your code. And the default behaviour of the API is still the same as before. However, we 
 extended the API, so **you can now choose the strategy of how to find the recommended points**.
 
-```http request
+```http
 POST /collections/{collection_name}/points/recommend
-
 {
   "positive": [100, 231],
   "negative": [718, [0.2, 0.3, 0.4, 0.5]],

qdrant-landing/content/articles/qdrant-1.3.x.md

@@ -69,7 +69,6 @@ Here is how you can configure the oversampling factor - define how many extra ve
 
 ```http
 POST /collections/{collection_name}/points/search
-
 {
   "params": {
     "quantization": {
@@ -122,7 +121,6 @@ When using the grouping API, add the `with_lookup` parameter to bring the inform
 
 ```http
 POST /collections/chunks/points/search/groups
-
 {
     // Same as in the regular search API
     "vector": [1.1],