Merge pull request #10 from edgararuiz/updates

Updates

.Rbuildignore

@@ -8,4 +8,6 @@ utils/
 ^pkgdown$
 ^\.github$
 ^codecov\.yml$
-^_mall_cache/
+^_mall_cache/
+^_readme_cache/
+^_prompt_cache/

.github/workflows/R-CMD-check.yaml

@@ -43,9 +43,16 @@ jobs:
 
       - uses: r-lib/actions/setup-r-dependencies@v2
         with:
-          extra-packages: any::rcmdcheck
-          needs: check
-
+          extra-packages: |
+            any::rcmdcheck
+          needs: |
+            check
+
+      - name: Installing dbplyr
+        run: |
+          pak::pak("dbplyr")
+        shell: Rscript {0}            
+
       - uses: r-lib/actions/check-r-package@v2
         with:
           upload-snapshots: true

.github/workflows/test-coverage.yaml

@@ -25,7 +25,10 @@ jobs:
 
       - uses: r-lib/actions/setup-r-dependencies@v2
         with:
-          extra-packages: any::covr, any::xml2
+          extra-packages: |
+            any::covr
+            any::xml2
+            any::dbplyr
           needs: coverage
 
       - name: Test coverage

DESCRIPTION

@@ -21,6 +21,7 @@ Imports:
     ollamar,
     rlang
 Suggests: 
+    dbplyr,
     testthat (>= 3.0.0)
 Config/testthat/edition: 3
 URL: https://edgararuiz.github.io/mall/

R/llm-extract.R

@@ -77,14 +77,10 @@ llm_extract.data.frame <- function(.data,
 llm_vec_extract <- function(x,
                             labels = c(),
                             additional_prompt = "") {
-  resp <- l_vec_prompt(
+  l_vec_prompt(
     x = x,
     prompt_label = "extract",
     labels = labels,
     additional_prompt = additional_prompt
   )
-  map_chr(
-    resp,
-    \(x) paste0(as.character(fromJSON(x, flatten = TRUE)), collapse = "|")
-  )
 }

R/m-backend-prompt.R

@@ -57,19 +57,20 @@ m_backend_prompt.mall_defaults <- function(backend, additional = "") {
       json_labels <- paste0("\"", labels, "\":your answer", collapse = ",")
       json_labels <- paste0("{{", json_labels, "}}")
       plural <- ifelse(no_labels > 1, "s", "")
+      text_multi <- ifelse(
+        no_labels > 1,
+        "Return the response in a simple list, pipe separated, and no headers. ",
+        ""
+      )
       list(
-        list(
-          role = "system",
-          content = "You only speak simple JSON. Do not write normal text."
-        ),
         list(
           role = "user",
           content = glue(paste(
             "You are a helpful text extraction engine.",
             "Extract the {col_labels} being referred to on the text.",
             "I expect {no_labels} item{plural} exactly.",
             "No capitalization. No explanations.",
-            "You will use this JSON this format exclusively: {json_labels} .",
+            "{text_multi}",
             "{additional}",
             "The answer is based on the following text:\n{{x}}"
           ))
@@ -100,7 +101,7 @@ l_vec_prompt <- function(x,
                          prompt = NULL,
                          ...) {
   # Initializes session LLM
-  backend <- llm_use(.silent = TRUE, force = FALSE)
+  backend <- llm_use(.silent = TRUE, .force = FALSE)
   # If there is no 'prompt', then assumes that we're looking for a
   # prompt label (sentiment, classify, etc) to set 'prompt'
   if (is.null(prompt)) {

R/m-backend-submit.R

@@ -47,6 +47,8 @@ m_backend_submit.mall_simulate_llm <- function(backend, x, prompt) {
     out <- map_chr(x, \(x) trimws(strsplit(x, "\\|")[[1]][[2]]))
   } else if (args$model == "echo") {
     out <- x
+  } else if (args$model == "prompt") {
+    out <- prompt
   }
   res <- NULL
   if (m_cache_use()) {
@@ -77,9 +79,6 @@ m_cache_record <- function(.args, .response, hash_args) {
 }
 
 m_cache_check <- function(hash_args) {
-  if (!m_cache_use()) {
-    return(invisible())
-  }
   folder_root <- m_cache_folder()
   resp <- suppressWarnings(
     try(read_json(m_cache_file(hash_args)), TRUE)

R/utils.R

@@ -1,13 +1,10 @@
-clean_names <- function(x, replace_periods = FALSE) {
+clean_names <- function(x) {
   x <- tolower(x)
   map_chr(
     x,
     \(x) {
       out <- str_replace_clean(x, " ")
       out <- str_replace_clean(out, "\\:")
-      if (replace_periods) {
-        out <- str_replace_clean(out, "\\.")
-      }
       out
     }
   )

tests/testthat/_snaps/llm-extract.md

@@ -1,3 +1,18 @@
+# Extract works
+
+    Code
+      llm_vec_extract("toaster", labels = "product")
+    Output
+      [[1]]
+      [[1]]$role
+      [1] "user"
+      
+      [[1]]$content
+      You are a helpful text extraction engine. Extract the product being referred to on the text. I expect 1 item exactly. No capitalization. No explanations.   The answer is based on the following text:
+      {x}
+      
+      
+
 # Extract on Ollama works
 
     Code
@@ -12,3 +27,10 @@
       2          laptop
       3 washing machine
 
+---
+
+    Code
+      llm_vec_extract("bob smith, 105 2nd street", c("name", "address"))
+    Output
+      [1] "bob smith | 105 2nd street"
+

tests/testthat/_snaps/llm-sentiment.md

@@ -1,3 +1,12 @@
+# Sentiment translates expected Spark SQL
+
+    Code
+      llm_sentiment(df_spark, x)
+    Output
+      <SQL>
+      SELECT `df`.*, ai_analyze_sentiment(`x`) AS `.sentiment`
+      FROM `df`
+
 # Sentiment on Ollama works
 
     Code

tests/testthat/_snaps/llm-summarize.md

@@ -1,3 +1,21 @@
+# Summarize translates expected Spark SQL
+
+    Code
+      llm_summarize(df_spark, x)
+    Output
+      <SQL>
+      SELECT `df`.*, ai_summarize(`x`, CAST(10.0 AS INT)) AS `.summary`
+      FROM `df`
+
+---
+
+    Code
+      llm_summarize(df_spark, x, max_words = 50)
+    Output
+      <SQL>
+      SELECT `df`.*, ai_summarize(`x`, CAST(50.0 AS INT)) AS `.summary`
+      FROM `df`
+
 # Summarize on Ollama works
 
     Code

tests/testthat/_snaps/zzz-cache.md

@@ -5,23 +5,24 @@
     Output
       _mall_cache/65
       _mall_cache/65/654ffcb598cbbdeb0c1c7ebd05239ed4.json
-      _mall_cache/65/65fd4ea24b687f30e12881a3c5ee4acc.json
       _mall_cache/72
       _mall_cache/72/72433745387c1c32b53ec05a77fa3b97.json
-      _mall_cache/73
-      _mall_cache/73/73962724ef856c31dffa1dfcae15daf1.json
       _mall_cache/76
       _mall_cache/76/768d2519acbc4cce84d8069f665ee0b2.json
       _mall_cache/7a
       _mall_cache/7a/7a500c559374fe68cf6cf57605d9de46.json
+      _mall_cache/7d
+      _mall_cache/7d/7d2911f2710df85c421e40c2991a6cac.json
+      _mall_cache/7e
+      _mall_cache/7e/7e20b6c382a19b95ea3e43d76af92b0f.json
       _mall_cache/83
       _mall_cache/83/83dafa73c6187bf2156f4f07d6fdfc36.json
-      _mall_cache/87
-      _mall_cache/87/8787f5fe54447691660ccb29bedbb420.json
       _mall_cache/ab
       _mall_cache/ab/abcd4e05a8a4cf61e0053a7a7bed1360.json
       _mall_cache/df
       _mall_cache/df/dfe963014849a2dbd09171246c45c511.json
+      _mall_cache/e6
+      _mall_cache/e6/e624d61d6a9dab8052ff6053b5be2f62.json
       _mall_cache/f4
       _mall_cache/f4/f47435d14c52e6a5c07400fb5f43db58.json
       _mall_cache/f7

tests/testthat/test-llm-extract.R

@@ -1,45 +1,44 @@
 test_that("Extract works", {
-  llm_use("simulate_llm", "echo", .silent = TRUE, .force = TRUE)
+  llm_use("simulate_llm", "prompt", .silent = TRUE, .force = TRUE)
 
-  expect_equal(
-    llm_vec_extract("{\"product\":\"toaster\"}", labels = "product"),
-    "toaster"
+  expect_snapshot(
+    llm_vec_extract("toaster", labels = "product")
   )
+})
+
+test_that("Extract data frame works", {
+  llm_use("simulate_llm", "echo", .silent = TRUE, .force = TRUE)
 
-  entries2 <- "{\"product\":\"toaster\", \"product\":\"TV\"}"
-  entries2_result <- "toaster|TV"
-  expect_equal(
-    llm_vec_extract(
-      entries2,
-      labels = "product"
-    ),
-    entries2_result
-  )
   expect_equal(
-    llm_extract(data.frame(x = entries2), x, labels = "product"),
-    data.frame(x = entries2, .extract = entries2_result)
+    llm_extract(data.frame(x = "test"), x, labels = "product"),
+    data.frame(x = "test", .extract = "test")
   )
+
   expect_equal(
     llm_extract(
-      .data = data.frame(x = entries2),
+      .data = data.frame(x = "test1|test2"),
       col = x,
       labels = c("product1", "product2"),
       expand_cols = TRUE
     ),
-    data.frame(x = entries2, product1 = "toaster", product2 = "TV")
+    data.frame(x = "test1|test2", product1 = "test1", product2 = "test2")
   )
+
   expect_equal(
     llm_extract(
-      .data = data.frame(x = entries2),
+      .data = data.frame(x = "test1|test2"),
       col = x,
       labels = c(y = "product1", z = "product2"),
       expand_cols = TRUE
     ),
-    data.frame(x = entries2, y = "toaster", z = "TV")
+    data.frame(x = "test1|test2", y = "test1", z = "test2")
   )
 })
 
 test_that("Extract on Ollama works", {
   skip_if_no_ollama()
   expect_snapshot(llm_extract(reviews_table(), review, "product"))
+  expect_snapshot(
+    llm_vec_extract("bob smith, 105 2nd street", c("name", "address"))
+  )
 })

tests/testthat/test-llm-sentiment.R

@@ -20,6 +20,13 @@ test_that("Sentiment works", {
   )
 })
 
+test_that("Sentiment translates expected Spark SQL", {
+  suppressPackageStartupMessages(library(dbplyr))
+  df <- data.frame(x = 1)
+  df_spark <- tbl_lazy(df, con = simulate_spark_sql())
+  expect_snapshot(llm_sentiment(df_spark, x))
+})
+
 test_that("Sentiment on Ollama works", {
   skip_if_no_ollama()
   vec_reviews <- reviews_vec()

tests/testthat/test-llm-summarize.R

@@ -21,6 +21,14 @@ test_that("Summarize works", {
   )
 })
 
+test_that("Summarize translates expected Spark SQL", {
+  suppressPackageStartupMessages(library(dbplyr))
+  df <- data.frame(x = 1)
+  df_spark <- tbl_lazy(df, con = simulate_spark_sql())
+  expect_snapshot(llm_summarize(df_spark, x))
+  expect_snapshot(llm_summarize(df_spark, x, max_words = 50))
+})
+
 test_that("Summarize on Ollama works", {
   skip_if_no_ollama()
   expect_snapshot(llm_summarize(reviews_table(), review, max_words = 5))

tests/testthat/test-llm-use.R

@@ -1,3 +1,15 @@
+test_that("Ollama not found error", {
+  local_mocked_bindings(
+    test_connection = function() {
+      x <- list()
+      x$status_code <- 400
+      x
+    }
+  )
+  .env_llm$defaults <- list()
+  expect_error(llm_use())
+})
+
 test_that("Init code is covered", {
   local_mocked_bindings(
     test_connection = function() {

tests/testthat/test-m-backend-prompt.R

@@ -0,0 +1,14 @@
+test_that("Prompt handles list()", {
+  llm_use(
+    backend = "simulate_llm",
+    model = "prompt",
+    .silent = TRUE,
+    .force = TRUE,
+    .cache = "_prompt_cache"
+  )
+  test_text <- "Custom:{prompt}\n{{x}}"
+  expect_equal(
+    llm_vec_custom(x = "new test", prompt = test_text),
+    list(list(role = "user", content = test_text))
+  )
+})

tests/testthat/test-m-backend-submit.R

@@ -8,3 +8,11 @@ test_that("Ollama code is covered", {
     "positive"
   )
 })
+
+test_that("No cache is saved if turned off", {
+  llm_use("simulate_llm", "echo", .silent = TRUE, .force = TRUE, .cache = "")
+  expect_equal(
+    llm_vec_custom("test text", "nothing new: "),
+    "test text"
+  )
+})

tests/testthat/test-zzz-cache.R

@@ -8,7 +8,7 @@ test_that("Ollama cache exists and delete", {
   skip_if_no_ollama()
   expect_equal(
     length(fs::dir_ls("_ollama_cache", recurse = TRUE)),
-    53
+    55
   )
   fs::dir_delete("_ollama_cache")
 })