Feat: add image preview to retrieval test. (#7610)

### What problem does this PR solve?

#7608

### Type of change

- [x] New Feature (non-breaking change which adds functionality)

rag/nlp/__init__.py

@@ -301,6 +301,7 @@ def tokenize_table(tbls, doc, eng, batch_size=10):
             d["content_with_weight"] = rows
             if img:
                 d["image"] = img
+                d["doc_type_kwd"] = "image"
             if poss:
                 add_positions(d, poss)
             res.append(d)
@@ -310,7 +311,9 @@ def tokenize_table(tbls, doc, eng, batch_size=10):
             d = copy.deepcopy(doc)
             r = de.join(rows[i:i + batch_size])
             tokenize(d, r, eng)
-            d["image"] = img
+            if img:
+                d["image"] = img
+                d["doc_type_kwd"] = "image"
             add_positions(d, poss)
             res.append(d)
     return res

rag/nlp/search.py

@@ -83,7 +83,7 @@ class Dealer:
         src = req.get("fields",
                       ["docnm_kwd", "content_ltks", "kb_id", "img_id", "title_tks", "important_kwd", "position_int",
                        "doc_id", "page_num_int", "top_int", "create_timestamp_flt", "knowledge_graph_kwd",
-                       "question_kwd", "question_tks",
+                       "question_kwd", "question_tks", "doc_type_kwd",
                        "available_int", "content_with_weight", PAGERANK_FLD, TAG_FLD])
         kwds = set([])
 
@@ -417,6 +417,7 @@ class Dealer:
                 "term_similarity": tsim[i],
                 "vector": chunk.get(vector_column, zero_vector),
                 "positions": position_int,
+                "doc_type_kwd": chunk.get("doc_type_kwd", "")
             }
             if highlight and sres.highlight:
                 if id in sres.highlight: