====== OnSight Analytics Interface ======

{{/dokuwiki/lib/images/smileys/icon_exclaim.gif|:!:}}**branch01 ist nur für die Zeit der Entwicklung die Heimat, keine Garantie für Uptime und Version**{{/dokuwiki/lib/images/smileys/icon_exclaim.gif|:!:}}

Host: **http://branch01.picalike.corpex-kunden.de:9095/**

Host: **http://branch01.picalike.corpex-kunden.de:9995/** (No timeout requests)(Please just one request per time)

Beispiel: **http://branch01.picalike.corpex-kunden.de:9095/hello**


===== git =====

Im git zu finden als:

<code>
$ git clone ssh://picalike@sg01.picalike.corpex-kunden.de/home/picalike/repositories/onsight_analytics.git
</code>

===== Allgemeines =====



===== MongoDB-Collections =====

<code>
attributes (alle picalike-Attribute)
categories (alle picalike-Kategorien)
genders (alle picalike-Genders)
</code>
**mit-Verwendung**

sessions


===== Apis =====

<code>
http://<host>:<port>/<api>
</code>

==== Solr Search ====



=== /update_solr_index (POST/json) ===

creates a new index in solr for shop_id or updates an existing one

**in:**

<code>
shop_id: <str>
session_history: <int>
</code>
Data used for update must have a session >= (latest_session - session_history)

**out:**

<code>
{}
</code>

=== /request_solr_index (POST/json) ===

make an search request to solr

**in:**

<code>
shop_id: <str> or <list of str>
query: <str>
start_date: yyyy-mm-dd <str>
end_date: yyyy-mm-dd <str>
start: <int, default:0> 
rows: <int, default:10>
sort: <str, optional>
</code>
example for start and row:


  * start: 5 means you will skip the first 5 results
  * rows: 5 means you will get 5 results

example for query_str:

<code>
gender:"Damen" AND name:"Schal" AND price:[* TO 2000]  
</code>
example for sort:

<code>
price desc
price asc
</code>
known field names for now: ''%%pid, price, name, text, brand, gender, images, picalike_gender, picalike_cat, shop_cat%%'' {{/dokuwiki/lib/images/smileys/icon_exclaim.gif|:!:}} some fields may be implemented in future

**out:**

<code>
{"message": <str>, "response": {"result_list": [<result_list>], "total_match": <int>, "stats":{"price": {"min": <int>, "max": <int>, "mean": <float>, "median": <int>}}}}
</code>

=== /is_synchronized (POST/json) ===

checks if session number is the same for import and solr

**in:**

<code>
{"shop_ids": [<str>]}
</code>
**out:**

<code>
{<shop_id>: <bool>}
</code>

==== Shops ====



=== /translate_id (POST) ===

translate apikey ↔ shop_id for feeds

**in:**

<code>
shop_id: <str>
OR
apikey: <str>
</code>
**out:**

<code>
{"message": <str>, "response": {<str> (apikey or shop_id)}}
</code>

=== /get_shops (GET) ===

returns all shop_ids (unique) of shops that have been crawled during last week (7*24*60*60 sec)

ohne cache, direkte Anfrage an die mongoDB

**in:**

<code>
{}
</code>
**in_opt:**

<code>
start_date=<str> (yyyy-mm-dd)
with_feed=<bool>
</code>
**out:**

<code>
{"message": <str>, "response": {"shop_ids": [<str>], "infos":{ <shop_id>: {"segment": <str>, "shop_name": <str>, "type": <str>}}}}
</code>

=== /get_shop_crawler_mapping (POST/json) ===

returns all crawler_shop_ids which are mapped to the shop_id

**in:**

<code>
shop_id: <str>
</code>
**out:**

<code>
{"message: <str>, "response": {"shop_ids": [<str>], "infos":{ <shop_id>: {"segment": <str>, "shop_name": <str>, "type": <str>}}}}
</code>

=== /remove_shop_crawler_mapping (POST/json) ===

remove all crawler_shop_ids which are mapped to the shop_id

**in:**

<code>
shop_id: <str>
</code>
**out:**

<code>
{"message: <str>, "response": {}}
</code>

=== /add_shop_crawler_mapping (POST/json) ===

update mapping of crawler_shop_ids to shop_id (will overwrite previous entries)

writes to shop_crawler_mapping in mongoDB

**in:**

<code>
shop_id: <str>
crawler_shops: [<str>]
</code>
**out:**

<code>
{"message: <str>, "response": {}}
</code>

=== /add_inspiration_shops (POST/json) ===

update mapping of inspiration_shop_ids to shop_id (will overwrite previous entries)

writes to shop_crawler_mapping in mongoDB

**in:**

<code>
shop_id: <str>
inspiration_shops: [<str>]  # list of shop_ids
</code>
**out:**

<code>
{"message: <str>, "response": {}}
</code>

=== /get_inspiration_shops (POST/json) ===

returns all inspiration_shop_ids which are mapped to the shop_id

**in:**

<code>
shop_id: <str>
</code>
**out:**

<code>
{"message: <str>, "response": {"shop_ids": [<str>], "infos":{ <shop_id>: {"segment": <str>, "shop_name": <str>, "type": <str>}}}}
</code>

=== /get_shop_stats (POST/json) ===

returns stats for a shop

**in:**

<code>
shop_ids: [<str>]
include_unmapped: bool,
max_age: <int, days, default: 30>
</code>
**out:**

<code>
{"message: <str>, "response": {
  "total": <int>  # total number of mapped products
  "by_cat": <dict>  # picalike_cat -> (absolute count, percent)
  # only if include_unmapped == True
  "unmapped_count": <int>
  "unmapped": <dict>  # <str, shop_cat> -> <int, absolute count>
}}
</code>

==== Products ====



=== /get_product_history (POST/json) ===

returns history for given products

**in:**

<code>
picalike_ids: [<str>]
</code>
**out:**

<code>
{"message": <str>, "response": {
  <picalike_id_1>: [{
    "timestamp": <float>, 
    "price": <int>, 
    "sort_key": <int>, 
    "position": [{"cat": [<str>], "pos": <float>}, ...]}, ...], 
  <picalike_id_2>: [], ...}
</code>
for each picalike_id, the array is sorted by “timestamp” price is in cents (int)


==== Diverse ====



=== /get_attributes (GET) ===

returns all picalike attributes

ohne cache, direkte Anfrage an die mongoDB

**in:**

<code>
{}
</code>
**out:**

<code>
{"message": <str>, "response": {"attributes": [<str>]}}
</code>

==== Categories ====



=== /get_categories (GET) ===

returns all picalike categories

ohne cache, direkte Anfrage an die mongoDB

**in:**

<code>
{}
</code>
**out:**

<code>
{"message": <str>, "response": {"categories": [<str>]}}
</code>

=== /get_shop_categories (GET) ===

return all categorie pathes from shop from last session

erstmalige Anfrage mongoDB, danach cache, update des caches mit Parameter use_cache=false

**in:**

<code>
shop_id=<str>
</code>
**in_opt:**

<code>
use_cache=false (default is true)
session_history=<int, default=1>
</code>
Shown data must haven an session >= (latest_session - session_history)

**out:**

<code>
{"message": <str>, "response": {"categories": [ list(<str>) ]}}
</code>
list(<str>) liefert die Baumstruktur der Kategorien des Shops, z.B. [ “Damen”, “Hosen”, “Jeans” ]


=== /add_category_relation (POST/json) ===

add a shop_category → picalike_category relation

**in:**

<code>
shop_id: <str>
p_cat: <picalike category str>
shop_cat: <shop category object>
</code>
siehe auch /get_categories und /get_shop_categories

**out:**


=== /remove_category_relation (POST/json) ===

remove a shop_category → picalike_category relation

**in:**

<code>
shop_id: <str>
shop_cat: <shop category object>
</code>
**out:**


=== /get_category_relation (GET) ===

get all relations for shop_id

**in:**

<code>
shop_id=<str>
</code>
**out:**

<code>
{"message": <str>, "response": [ tupel(<picalike category str>, <shop category object>) ]}
</code>
Das Tupel wird in json als eine Liste mit zwei Elementen abgelegt.


=== /get_top_categories (GET) ===

get top categories for all shops in shop_id together

**in:**

<code>
shop_id=[<str>]
</code>
**in_opt:**

<code>
top_n=<int>
</code>
**out:**

<code>
{"message": <str>, "response": {"n_products": <int>, "top_cats": [ tuple(<picalike category str>, <number of products int>) ]}}
</code>
Categories are ordered from most to least frequent.


=== /get_categorization (POST) ===

get picalike category suggestions for origin shop category

**in:**

<code>
shop_id:<str>
</code>
**in_opt:**

<code>
shop_cat:[<str>, <str>,...] # optional: default looks for all existing shop categories
</code>
**out:**

<code>
{"message": <str>, "response": [{"shop_cat": "Damen_Hosen_Lange Hosen", "picalike_cat": "fashion_leg_pants", "rel_freq": 1.0}]}
</code>

=== http:%%//%%branch01.picalike.corpex-kunden.de:5003/prepare_categorization (GET) ===

prepares suggestions of picalike categories for origin shop categories based on the relative frequencies of the picalike categories in an origin shop category. Saves results in MongoDB:

Sandy Picalike –> patrick_test –> categorization


==== Brands ====



=== /get_shop_brands (POST) ===

return all brands from shop from last session

**in:**

<code>
shop_id:<str>
</code>
**out:**

<code>
{"message": <str>, "response": {"brands": [ <str> ], "count":{<brand>: <int>}}}
</code>

==== Genders ====



=== /get_genders (GET) ===

returns all picalike genders

ohne cache, direkte Anfrage an die mongoDB

**in:**

**out:**

<code>
{"message": <str>, "response": {"genders": [<str>]}}
</code>

=== /get_shop_genders (GET) ===

return all genders from shop from last session

erstmalige Anfrage mongoDB, danach cache, update des caches mit Parameter use_cache=false

**in:**

<code>
shop_id=<str>
</code>
**in_opt:**

<code>
use_cache=false (default is true)
</code>
**out:**

<code>
{"message": <str>, "response": {"genders": [ <str> ]}}
</code>

=== /add_gender_relation (POST/json) ===

add a shop_genders → picalike_genders relation

**in:**

<code>
shop_id: <str>
p_gender: <picalike genders str>
shop_gender: <shop genders str>
</code>
**out:**


=== /remove_gender_relation (POST/json) ===

remove a shop_genders → picalike_genders relation

**in:**

<code>
shop_id: <str>
shop_gender: <shop genders str>
</code>
**out:**


=== /get_gender_relation (GET) ===

get all relations for shop_id

**in:**

<code>
shop_id=<str>
</code>
**out:**

<code>
{"message": <str>, "response": [ tupel(<picalike genders str>, <shop genders str>) ]}
</code>
Das Tupel wird in json als eine Liste mit zwei Elementen abgelegt.


==== Reports ====



=== /existing_report (GET) ===

checks if report name exists

**in:**

<code>
report_name=<str>
shop_id=<str>
</code>
**out:**

<code>
{"message": <str>, "response": {"exists": <bool>}}
</code>

=== /list_reports (GET) ===

return all reports for given product and shop

**in:**

<code>
shop_id=<str>
product_id=<str>
</code>
**out:**

<code>
{"message": <str>, "response": [{"report_id": <str>, "user_id": <str>, "report_name": <str>}]}
</code>

=== /get_all_reports (POST) ===

return all reports that match the filters (only manually created reports)

**in:**

<code>
ALL OPTIONAL
shop_id: <str>
user_id: <str>
product_id: <str>
date_from: <str, "%m/%d/%Y">  # e.g.: "01/31/2019"
date_till: <str, "%m/%d/%Y">
</code>
**out:**

<code>
{"message": <str>, "response": [{"report_id": <str>, "user_id": <str>, "report_name": <str>, ...}]}
</code>

=== /add_report (POST/json) ===

add a new report or edit if report ID is provided

**in:**

<code>
shop_id: <str>
product_id: <str>
user_id: <str>
report_name: <str>
filter: { see below }
</code>
**in opt:**

<code>
report_id: <str>
</code>
**out:**

<code>
{"message": <str>, "response": {"report_id": <str>}
</code>
**filter** explanation:

<code>
cluster_dist: <float> - similarity
cluster_price_from <str> - e.g. "10,00"
cluster_price_till <str> - e.g. "29,99"
cluster_trendsetters <list of base64 encoded shop_ids>
cluster_competitors <list of base64 encoded shop_ids>
cluster_brands <list of brand names>
cluster_genders <list of picalike gender names>
cluster_date_range <str> - e.g. "01/13/2019 - 01/17/2019"
cluster_categories <list of picalike category names>
</code>
{{/dokuwiki/lib/images/smileys/icon_exclaim.gif|:!:}} cluster_genders can contain the value “all” which needs special treatment


=== /remove_report (POST/json) ===

remove a report

**in:**

<code>
report_id: <str>
</code>
**out:**

<code>
{"message": <str>}
</code>

=== /get_report (GET) ===

get a report

**in:**

<code>
report_id=<str>
</code>
**out:**

<code>
{"message": <str>, "response": {"report_id": <str>, "report_name": <str>, "product_id": <str>, "shop_id": <str>, "user_id": <str>, "filter": {}, "date": <date>}}
</code>

=== /exclude_product (POST) ===

remove product from results

**in:**

<code>
{"report_id": <str>, "picalike_id": <str>, "user_id": <str>}
</code>

=== /unexclude_product (POST) ===

remove product from exclude list

**in:**

<code>
{"report_id": <str>, "picalike_id": <str>}
</code>

=== /get_excluded (GET) ===

get list of excluded products

**in:**

<code>
report_id=<str>
</code>
**out:**

<code>
{"message": <str>, "response": {"excluded": [(<str: picalike_id>, <str: user_id>)]}}
</code>

==== Cluster ====



=== /comp_cluster (POST/json) ===

creates all the data that is needed by comp_cluster.php. It is also used to create automatic reports

**in:**

<code>
shop_id: <str>
prod_id: <str> or image: <base64 encoded image bytes>
optional:
report_id: <str>, if given, the report is stored immediately
price: <int, price in cents>  # only needed if image is given
start: <int, default: 0>
rows: <int, default: 10>
limit: <int, default: 500>  -- used for nearest neighbor search
filter fields with the same name and in the same format as in /add_report 
</code>
**out:**

<code>
shop_id: <str>
prod_id: <str>  # None if image was supplied
reco_data: <dict of data that is given to the reco algorithm> currently: { min_price, max_price, mean_price, ref_price }
reco: { timestamp, error, uncertain, huge_markup, price_range, margin_range }
cluster_trend: <int>
excluded: <list of tuples, (picalike_ids, user_id)>
total_match: <int> - total number of solr matches
cluster: <list of cluster products + distance, prod_trend and prod_history> 
</code>
if **report_id** is supplied, the report will be saved on the fly

**reco_data** is a dictionary with the required data to calculate the recommendation. Currently the fields //min_price, max_price, mean_price, ref_price// are needed.

**reco** contains at least //timestamp// and //error//. //error// can be either a boolean or a string. //uncertain// is a boolean that is set to True if the algorithm does not want to give a recommendation. //huge_markup// is a boolean that is set to True if the price of the reference article is significantly below the mean price in the cluster. //price_range// returns a list of prices [low, high]. //margin_range// is the relative price increase in percent [low, high]

**cluster_trend** indicates whether the products in the cluster of similiar products from trendsetter shops have recently been more (close to 1) or less (near -1) popular. A value of 0 indicates that there was no change in popularity in recently. The value 1 indicates a fast growing popularity, the value 0 predicts a negativ trend.

**cluster** contains a list of similar products. The documents are the solr responses with the additional fields: //distance//, //prod_trend// and //prod_history//. The reference product is always at the beginning of the list. Data generated from similiar products from shop competitors.


=== /get_cluster_report (POST/json) ===

returns the most recent results of the batch processing of the cluster reports

**in:**

<code>
report_id: <str>  - can be a report_id or a picalike_id
</code>
**out:**

<code>
data: <dict> - a merge of 'reco_data' and 'reco' from /comp_cluster
</code>

==== Trends ====



=== http:%%//%%branch01.picalike.corpex-kunden.de:5002/prepare_trends (GET) ===

ONLY USE DAILY - PROCESS TAKES MORE THAN 30 MIN Receives a get requests and downloads the product infos from solr and from the API “http://branch01.picalike.corpex-kunden.de:9095/get_product_history” to find trendy products. It then forecast the future trend with double double_exponential_smoothing and saves its results in the MongoDB


=== /sort_trend_products (POST/json) ===

returns the most trendy products from the recent trend analysis from the MongoDB only looks into products from shops, that are mapped for the specific shop

**in:**

<code>
shop_id: <str> # only feed shops accepted
optional:
limit:<int> # default: 10
category: [<str>] # default: []
gender: [<str>] # default: ["Men", "Women", "undefined"] --> some products do not have gender info
selected_shops:[<str>] # default: [] --> looks for all shops
</code>
**out:**

<code>
{"message": <str>, "response": [{"product_id":<str>, "category":<str>,"datetime":<datetime>,"gender":<str>,"pos_mov":<float>,"shop_id":<str>}]
</code>

=== /sort_trend_brands (POST/json) ===

returns the most trendy brands from the recent trend analysis from the MongoDB only looks into products from shops, that are mapped for the specific shop

**in:**

<code>
shop_id: <str> # only feed shops accepted
optional:
limit:<int> # default: 10
category: [<str>] # default: []
gender: [<str>] # default: ["Men", "Women", "undefined"] --> some products do not have gender info
selected_shops:[<str>] # default: [] --> looks for all shops
</code>
**out:**

<code>
{"message": <str>, "response": [{"brand":<str>, "category":[<str>],"datetime":<datetime>,"gender":[<str>],"pos_mov":<float>,"shop_id":[<str>]}]  
</code>

=== /sort_trend_category (POST/json) ===

returns the most trendy categories from the recent trend analysis from the MongoDB only looks into products from shops, that are mapped for the specific shop

**in:**

<code>
shop_id: <str> # only feed shops accepted
optional:
limit:<int> # default: 10
gender: [<str>] # default: ["Men", "Women", "undefined"] --> some products do not have gender info
selected_shops:[<str>] # default: [] --> looks for all shops
</code>
**out:**

<code>
{"message": <str>, "response": [{"category":<str>, "datetime":<datetime>","gender":[<str>],"pos_mov":<float>,"shop_id":[<str>]}]  
</code>

=== /sort_trend_colors (POST/json) ===

returns the most trendy colors from the recent trend analysis from the MongoDB only looks into products from shops, that are mapped for the specific shop

**in:**

<code>
shop_id: <str> # only feed shops accepted
optional:
limit:<int> # default: 10
category: [<str>] # default: []
gender: [<str>] # default: ["Men", "Women", "undefined"] --> some products do not have gender info
selected_shops:[<str>] # default: [] --> looks for all shops
</code>
**out:**

<code>
{"message": <str>, "response": [{"color":<str>, "category":[<str>],"datetime":<datetime>,"gender":[<str>],"pos_mov":<float>,"shop_id":[<str>]}]   
</code>

=== /sort_trend_attributes (POST/json) ===

returns the most trendy attributes from the recent trend analysis from the MongoDB only looks into products from shops, that are mapped for the specific shop

**in:**

<code>
shop_id: <str> # only feed shops accepted
optional:
limit:<int> # default: 10
category: [<str>] # default: []
gender: [<str>] # default: ["Men", "Women", "undefined"] --> some products do not have gender info
selected_shops:[<str>] # default: [] --> looks for all shops
</code>
**out:**

<code>
{"message": <str>, "response": [{"attribute":<str>, "category":[<str>],"datetime":<datetime>,"gender":[<str>],"pos_mov":<float>,"shop_id":[<str>]}] 
</code>

=== /get_sim_trend_products (POST/json) ===

looks for products in feed shop that are similiar to the given trend product API is currently used by /sort_sim_trend_products

**in:**

<code>
{"shop_id_feed":<str>,
"shop_id_crawler":<str>,
"product_id":<int>,
"distance_max": <float> # optional: limits similiarity distance, default is 0.5
}
</code>
**out:**

<code>
{"message":"OK","response":{
            "product_ids:["96607881#bonprix_de_feed,
                          "91524296#bonprix_de_feed",
                          "90838081#bonprix_de_feed",
                          "95056195#bonprix_de_feed",
                          "95020895#bonprix_de_feed"]
            }
 }
</code>

=== http:%%//%%frontend01-hpc.picalike.corpex-kunden.de:5003/get_cat_trends (GET) Port 5003 ===

Returns the most trendy products from a category of one shop. Only returns available products. Products sorted in descending order by cluster_trend.

**in:**

<code>
shop_id: <str>
prod_id: <int> # Becomes optional, if "i" as prod_ref is given as argument
optional:
limit:<int> # default: 10
brand: <str> or [<str>, <str>,...] # default: all
shop_cat: <str> or [<str>, <str>,...] # default: all
gender: <str> or [<str>, <str>,...] # default: all
size: <int> or [<int>, <int>,...] # CURRENTLY NOT POSSIBLE TO FILTER. default: all
max_price: <int> # default: no max
min_price: <int> # default: no min 
category_limitation: True/False # default: True
i: <picalike_id> # if "i" is set, gender and category are same as given picalike_id. Default: None
</code>
**out:**

<code>
{"count": 10,
"description": "Category Trends",
"generator": "http://picalike.com",
"modified": "2019-04-29 08:24:04.480205",
"title": "picalike Request",
"ids": {"0": {"brand": "NO", "cluster_trend": 0.0, "extraimg": "https://www.witt.eu/product/resized/029/029.00K3F.072-127.002.u_5.jpg", "gender": "Damenmode", "shop_cat":"Frauen_Bekleidung_Hosen_Sweatpants", "id": "184097", "img": "https://www.witt.eu/product/resized/027/027.00KAT.022-123.013.i_5.jpg", "location": "https://www.witt-weiden.de/287202?articleNumber=184097", "name": "Hose", "price": 1599},
"1":{...}}
}
</code>

==== Krawla ====



=== http:%%//%%branch01.picalike.corpex-kunden.de:9095/get_krawla_sum (GET) ===

**in:**

<code>
shop_id: <str> # only feed shops accepted
</code>
**out:**

<code>
{"message": <str>, 
"response": {"w1": {"total_all" : <int>, "total_new_prod" : <int>, "total_sale" : <int> },
"w2": {"total_all" : <int>, "total_new_prod" : <int>, "total_sale" : <int> },
"w3": {"total_all" : <int>, "total_new_prod" : <int>, "total_sale" : <int> },
"w4": {"total_all" : <int>, "total_new_prod" : <int>, "total_sale" : <int> }, 
"w_all" : {"total_all" : <int>, "total_new_prod" : <int>, "total_sale" : <int> }
}
}
</code>